id int64 1 119 | dataset stringclasses 1
value | instruction stringlengths 69 141 | input stringlengths 0 832 | output stringlengths 15 8.9k | state unknown | question_block unknown | answer_block unknown | question_count int64 0 9 | has_answer bool 2
classes | model stringclasses 1
value | function_type stringclasses 3
values | doc_title stringclasses 22
values | section stringlengths 4 57 | category stringclasses 5
values | source stringclasses 22
values | tags listlengths 5 10 | credit stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | INSTRUCT_JEV | Explain 'Demos' from the TypeSafe AI Noul documentation for System One (Jev). | Interactive examples showing what's possible with TypeSafe. | # Demos
> Interactive examples showing what's possible with TypeSafe. | null | null | null | 0 | false | jev-latest | noul | Demos | Demos | _demos | _demos/_smart_home_assistant.md | [
"typesafe",
"jev",
"system-one",
"noul",
"demos",
"examples",
"interactive",
"possible",
"showing"
] | TypeSafe AI - https://typesafe.ai |
2 | INSTRUCT_JEV | Explain 'Available demos' from the TypeSafe AI Noul documentation for System One (Jev). | Interactive examples showing what's possible with TypeSafe. | ## Available demos
* [Smart Home Assistant Demo](/demos/smart-home) - Evaluate user smart home requests with speculative questions and LLM fallback.
<Tip>
We're always keen to learn how people are making use of our primitives. If you've found a killer use case you think should be mentioned here, feel free to drop u... | null | null | null | 0 | false | jev-latest | noul | Demos | Available demos | _demos | _demos/_smart_home_assistant.md | [
"typesafe",
"jev",
"system-one",
"noul",
"demos",
"available",
"home",
"smart",
"always",
"assistant"
] | TypeSafe AI - https://typesafe.ai |
3 | INSTRUCT_JEV | Explain 'Smart home assistant demo' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | # Smart home assistant demo
> Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | Smart home assistant demo | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"assistant",
"demo",
"home",
"smart",
"code",
"evaluate"
] | TypeSafe AI - https://typesafe.ai |
4 | INSTRUCT_JEV | Explain 'Check it out in action' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | ## Check it out in action
<Frame>
<iframe src="https://www.loom.com/embed/18c4dbcf8db546dfb2d7f2ef018e78e4" title="Smart home assistant demo video" allow="fullscreen; picture-in-picture" style={{ width: '100%', aspectRatio: '16 / 9', border: 'none' }} />
</Frame> | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | Check it out in action | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"action",
"check",
"frame",
"allow",
"aspectratio",
"assistant"
] | TypeSafe AI - https://typesafe.ai |
5 | INSTRUCT_JEV | Explain 'How it works' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | ## How it works | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | How it works | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"works"
] | TypeSafe AI - https://typesafe.ai |
6 | INSTRUCT_JEV | Explain 'Speculative fan-out' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | ### Speculative fan-out
The chief pattern demonstrated here is [speculative fan-out](/patterns/fan-out). Each user request is evaluated against a long list of questions, including many that will end up irrelevant for most requests.
Let's consider the following user request:
> "Turn off all of the lights in the house... | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | Speculative fan-out | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"request",
"user",
"fan-out",
"lights",
"speculative",
"before"
] | TypeSafe AI - https://typesafe.ai |
7 | INSTRUCT_JEV | Explain 'The wrong way: sequential API calls' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | #### The wrong way: sequential API calls
The wrong way to do this would be to separate the questions in to multiple API calls, waiting to ask questions only once you are certain you need the answer:
* "What category of request is this?" (smarthome command)
Then, only once you know it's a smarthome command:
* "What ... | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | The wrong way: sequential API calls | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"calls",
"lights",
"once",
"only",
"request",
"targeting"
] | TypeSafe AI - https://typesafe.ai |
8 | INSTRUCT_JEV | Explain 'TypeSafe and LLM pairing' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | ### TypeSafe and LLM pairing
This demo also shows how TypeSafe can be paired with LLMs to handle a system that sometimes requires a string-generation step:
**Splitting a compound user request:** One of the questions in this demo is a Noul question identifying if the user request is asking for more than one distinct a... | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | TypeSafe and LLM pairing | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"system",
"request",
"response",
"user",
"demo",
"fast"
] | TypeSafe AI - https://typesafe.ai |
9 | INSTRUCT_JEV | Explain 'Run it yourself' from the TypeSafe AI Noul documentation for System One (Jev). | Demo code: a smart home assistant that uses TypeSafe to evaluate user requests. | ## Run it yourself
This demo is a simple Vite/React single-page app that uses the TypeSafe API to evaluate user requests. The full source code will be available on GitHub at release. Its README includes instructions for running the demo locally and an overview of which bits of the source code are responsible for which... | null | null | null | 0 | false | jev-latest | noul | Smart home assistant demo | Run it yourself | _demos | _demos/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"demo",
"code",
"source",
"yourself",
"available",
"bits"
] | TypeSafe AI - https://typesafe.ai |
10 | INSTRUCT_JEV | Define the TypeSafe System One Score question(s) used for 'Step 1: score each dimension independently'. | {"python_depth": {"type": "score", "instructions": "How much depth of python experience does this candidate have, based on the supplied resume?", "criteria": ["No Python experience mentioned", "Mentioned but no detail", "Used in projects, some specifics", "Primary language, multiple projects", "Deep expertise: architec... | "" | {
"python_depth": {
"type": "score",
"instructions": "How much depth of python experience does this candidate have, based on the supplied resume?",
"criteria": [
"No Python experience mentioned",
"Mentioned but no detail",
"Used in projects, some specifics",
"Primary language, mult... | null | 4 | false | jev-latest | score | Composite scoring | Step 1: score each dimension independently | _patterns | _patterns/_composite_scoring.md | [
"typesafe",
"jev",
"system-one",
"score",
"experience",
"mentioned",
"candidate",
"criteria",
"instructions",
"much"
] | TypeSafe AI - https://typesafe.ai | |
11 | INSTRUCT_JEV | Explain 'Composite scoring' from the TypeSafe AI Score documentation for System One (Jev). | Break a complex judgment into atomic scores, combine with weights you control in code. | # Composite scoring
> Break a complex judgment into atomic scores, combine with weights you control in code.
Oftentimes we want to rank a set of items based on several criteria at once. Composite scoring is an easy way to think about this: break the judgment into independent dimensions, score each one separately, and... | null | null | null | 0 | false | jev-latest | score | Composite scoring | Composite scoring | _patterns | _patterns/_composite_scoring.md | [
"typesafe",
"jev",
"system-one",
"score",
"composite",
"scoring",
"break",
"code",
"combine",
"control"
] | TypeSafe AI - https://typesafe.ai |
12 | INSTRUCT_JEV | Explain 'Example: resume screening' from the TypeSafe AI Noul documentation for System One (Jev). | Break a complex judgment into atomic scores, combine with weights you control in code. | ## Example: resume screening
Let's imagine you are processing resumes for engineering roles. You want to rank the candidates based on several criteria, and ultimately select the top X candidates for further review. | null | null | null | 0 | false | jev-latest | noul | Composite scoring | Example: resume screening | _patterns | _patterns/_composite_scoring.md | [
"typesafe",
"jev",
"system-one",
"noul",
"candidates",
"example",
"resume",
"screening",
"based",
"criteria"
] | TypeSafe AI - https://typesafe.ai |
13 | INSTRUCT_JEV | Explain 'Step 2: combine with weights' from the TypeSafe AI Score documentation for System One (Jev). | Break a complex judgment into atomic scores, combine with weights you control in code. | ### Step 2: combine with weights
Each dimension is normalized to 0–1 and weighted. The weights give you an easy way to adjust the relative importance of each dimension, without losing any of the nuance of the individual scores.
```python title="scoring.py" theme={null}
py = response.answers["python_depth"].score... | null | null | null | 0 | false | jev-latest | score | Composite scoring | Step 2: combine with weights | _patterns | _patterns/_composite_scoring.md | [
"typesafe",
"jev",
"system-one",
"score",
"answers",
"response",
"weights",
"arch",
"general",
"lead"
] | TypeSafe AI - https://typesafe.ai |
14 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Step 1: determine the user's intent'. | {"intent": {"type": "choice", "instructions": "What action is the user requesting?", "criteria": {"check_balance": "Check the balance of an account", "approve_transfer": "Approve the pending transfer request", "other": "Something else"}}} | "" | {
"intent": {
"type": "choice",
"instructions": "What action is the user requesting?",
"criteria": {
"check_balance": "Check the balance of an account",
"approve_transfer": "Approve the pending transfer request",
"other": "Something else"
}
}
} | null | 1 | false | jev-latest | choice | Confidence-gated routing | Step 1: determine the user's intent | _patterns | _patterns/_confidence-gated_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"intent",
"user",
"approve",
"balance",
"check",
"determine"
] | TypeSafe AI - https://typesafe.ai | |
15 | INSTRUCT_JEV | Explain 'Confidence-gated routing' from the TypeSafe AI Choice documentation for System One (Jev). | Use confidence as a second axis. The answer tells you what; confidence tells you whether to act. | # Confidence-gated routing
> Use confidence as a second axis. The answer tells you what; confidence tells you whether to act.
One of TypeSafe's most powerful features is [confidence](/confidence). By being intentional with the way you gate decisions on confidence, you can build systems that are both reliable and safe... | null | null | null | 0 | false | jev-latest | choice | Confidence-gated routing | Confidence-gated routing | _patterns | _patterns/_confidence-gated_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"confidence-gated",
"routing",
"tells",
"answer",
"axis"
] | TypeSafe AI - https://typesafe.ai |
16 | INSTRUCT_JEV | Explain 'Example: voice banking commands' from the TypeSafe AI Noul documentation for System One (Jev). | Use confidence as a second axis. The answer tells you what; confidence tells you whether to act. | ## Example: voice banking commands
Let's imagine you are building a voice banking interface to allow the user to interact with their account verbally. While you always want to have reasonable confidence in interpreting the user's intent, some actions are riskier than others and thus demand a higher confidence threshol... | null | null | null | 0 | false | jev-latest | noul | Confidence-gated routing | Example: voice banking commands | _patterns | _patterns/_confidence-gated_routing.md | [
"typesafe",
"jev",
"system-one",
"noul",
"banking",
"voice",
"commands",
"confidence",
"example",
"user"
] | TypeSafe AI - https://typesafe.ai |
17 | INSTRUCT_JEV | Explain 'Step 2: confidence-gated routing' from the TypeSafe AI Choice documentation for System One (Jev). | Use confidence as a second axis. The answer tells you what; confidence tells you whether to act. | ### Step 2: confidence-gated routing
```python theme={null}
action = response.answers["intent"]
# Below 0.6 confidence on any action, route to a human
if action.confidence < 0.6:
route_to_support_agent(account_id)
elif action.choice == "check_balance":
# Low stakes. 0.6 confidence is sufficient.
show_bal... | null | null | null | 0 | false | jev-latest | choice | Confidence-gated routing | Step 2: confidence-gated routing | _patterns | _patterns/_confidence-gated_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"action",
"account",
"balance",
"high",
"transfer"
] | TypeSafe AI - https://typesafe.ai |
18 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Step 1: classify intent and complexity'. | {"intent": {"type": "choice", "instructions": "The primary intent of this customer message", "criteria": {"order_status": "Asking about an existing order", "product_question": "Asking about a product before buying", "return_exchange": "Wants to return or exchange something", "complaint": "Unhappy with experience, wants... | "" | {
"intent": {
"type": "choice",
"instructions": "The primary intent of this customer message",
"criteria": {
"order_status": "Asking about an existing order",
"product_question": "Asking about a product before buying",
"return_exchange": "Wants to return or exchange something",
"co... | null | 2 | false | jev-latest | choice | Intent routing | Step 1: classify intent and complexity | _patterns | _patterns/_intent_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"intent",
"complexity",
"about",
"asking",
"classify",
"criteria"
] | TypeSafe AI - https://typesafe.ai | |
19 | INSTRUCT_JEV | Explain 'Intent routing' from the TypeSafe AI Choice documentation for System One (Jev). | Classify incoming requests and route each to the optimal handler: deterministic logic, a specialist LLM, or a human. | # Intent routing
> Classify incoming requests and route each to the optimal handler: deterministic logic, a specialist LLM, or a human.
Not every user request needs the same kind of handler. Some can be answered with a database lookup. Some need an LLM with domain-specific context. Some need a human. TypeSafe can sit... | null | null | null | 0 | false | jev-latest | choice | Intent routing | Intent routing | _patterns | _patterns/_intent_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"handler",
"some",
"human",
"intent",
"need",
"routing"
] | TypeSafe AI - https://typesafe.ai |
20 | INSTRUCT_JEV | Explain 'Example: customer service routing' from the TypeSafe AI Choice documentation for System One (Jev). | Classify incoming requests and route each to the optimal handler: deterministic logic, a specialist LLM, or a human. | ## Example: customer service routing
Let's imagine you are building a customer service system. Messages come in and need to be routed to the right handler. Rather than sending every message through an expensive LLM to figure out what kind of request it is, you classify first and route accordingly. | null | null | null | 0 | false | jev-latest | choice | Intent routing | Example: customer service routing | _patterns | _patterns/_intent_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"customer",
"service",
"example",
"routing",
"accordingly",
"building"
] | TypeSafe AI - https://typesafe.ai |
21 | INSTRUCT_JEV | Explain 'Step 2: route to the optimal handler' from the TypeSafe AI Choice documentation for System One (Jev). | Classify incoming requests and route each to the optimal handler: deterministic logic, a specialist LLM, or a human. | ### Step 2: route to the optimal handler
```python title="routing.py" theme={null}
def route_ticket(ticket_id, response):
intent = response.answers["intent"]
complexity = response.answers["complexity"]
if intent.confidence < 0.5:
# If we don't have enough confidence to classify, route to a human a... | null | null | null | 0 | false | jev-latest | choice | Intent routing | Step 2: route to the optimal handler | _patterns | _patterns/_intent_routing.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"complexity",
"intent",
"route",
"ticket",
"human"
] | TypeSafe AI - https://typesafe.ai |
22 | INSTRUCT_JEV | Explain 'Patterns' from the TypeSafe AI Noul documentation for System One (Jev). | Architectural patterns for building systems with TypeSafe. | # Patterns
> Architectural patterns for building systems with TypeSafe.
TypeSafe is designed to sit within a larger system, powering decisions with AI. Learning to think in terms of discrete, atomic decisions that compose into complex system behavior is a key skill for getting the most out of TypeSafe.
This section ... | null | null | null | 0 | false | jev-latest | noul | Patterns | Patterns | _patterns | _patterns/_patterns.md | [
"typesafe",
"jev",
"system-one",
"noul",
"patterns",
"confidence",
"decisions",
"primitives",
"system",
"architectural"
] | TypeSafe AI - https://typesafe.ai |
23 | INSTRUCT_JEV | Explain 'The patterns' from the TypeSafe AI Choice documentation for System One (Jev). | Architectural patterns for building systems with TypeSafe. | ## The patterns
| Pattern | What it does | Benefits |
| -------------------------------------------------------- | ---------------------------------------------... | null | null | null | 0 | false | jev-latest | choice | Patterns | The patterns | _patterns | _patterns/_patterns.md | [
"typesafe",
"jev",
"system-one",
"choice",
"patterns",
"cost",
"speed",
"fan-out",
"intent",
"reliability"
] | TypeSafe AI - https://typesafe.ai |
24 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Step 1: speculative fan-out'. | Hi, I placed an order (#98423) last Thursday and was charged twice. I also can't log in after the site update, and adding Apple Pay would be really helpful. This is getting frustrating. | {"category": {"type": "choice", "instructions": "Determine the broad category of this support ticket", "criteria": {"bug_report": "The user is reporting something that is broken or producing errors", "billing": "Charges, invoices, refunds, subscriptions", "feature_request": "The user is requesting new functionality", "... | "Hi, I placed an order (#98423) last Thursday and was charged twice. I also can't log in after the site update, and adding Apple Pay would be really helpful. This is getting frustrating." | {
"category": {
"type": "choice",
"instructions": "Determine the broad category of this support ticket",
"criteria": {
"bug_report": "The user is reporting something that is broken or producing errors",
"billing": "Charges, invoices, refunds, subscriptions",
"feature_request": "The user ... | null | 5 | false | jev-latest | choice | Speculative fan-out | Step 1: speculative fan-out | _patterns | _patterns/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"choice",
"instructions",
"user",
"criteria",
"feature",
"issue",
"refund"
] | TypeSafe AI - https://typesafe.ai |
25 | INSTRUCT_JEV | Explain 'Speculative fan-out' from the TypeSafe AI Noul documentation for System One (Jev). | Send many questions in a single call, including speculative ones, and let your code decide what's relevant. | # Speculative fan-out
> Send many questions in a single call, including speculative ones, and let your code decide what's relevant.
Because TypeSafe supports sending many questions in a single API call, we recommend putting all of the questions your system needs in a single request, and then using code to decide what... | null | null | null | 0 | false | jev-latest | noul | Speculative fan-out | Speculative fan-out | _patterns | _patterns/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"noul",
"call",
"single",
"speculative",
"code",
"decide",
"fan-out"
] | TypeSafe AI - https://typesafe.ai |
26 | INSTRUCT_JEV | Explain 'Example: support ticket triage' from the TypeSafe AI Choice documentation for System One (Jev). | Send many questions in a single call, including speculative ones, and let your code decide what's relevant. | ## Example: support ticket triage
Let's imagine you are building a support system that needs to triage support tickets. You need to classify the ticket into a category. If it's a bug report, you also need to determine the severity of the bug.
Instead of asking for the category first and then the severity in a follow-... | null | null | null | 0 | false | jev-latest | choice | Speculative fan-out | Example: support ticket triage | _patterns | _patterns/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"choice",
"support",
"ticket",
"severity",
"triage",
"category",
"example"
] | TypeSafe AI - https://typesafe.ai |
27 | INSTRUCT_JEV | Explain 'Step 2: route with code' from the TypeSafe AI Choice documentation for System One (Jev). | Send many questions in a single call, including speculative ones, and let your code decide what's relevant. | ### Step 2: route with code
Your code decides what is relevant based on the classification result:
```python title="triage.py" theme={null}
category = response.answers["category"]
bug_severity = response.answers["bug_severity"]
bug_repro = response.answers["has_reproducible_steps"]
refund = response.answers["refund_r... | null | null | null | 0 | false | jev-latest | choice | Speculative fan-out | Step 2: route with code | _patterns | _patterns/_speculative_fan-out.md | [
"typesafe",
"jev",
"system-one",
"choice",
"category",
"response",
"ticket",
"answers",
"frustration",
"refund"
] | TypeSafe AI - https://typesafe.ai |
28 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Structured instructions'. | {"source_text": "Invoice #4471 issued March 3, 2026 to Beaver Dam Logistics for $12,840.00, net 30."} | {"invoice_number_is_correct": {"type": "noul", "instructions": {"field": {"name": "invoice_number", "type": "string", "description": "The identifier printed on the invoice."}, "extracted_value": "4471", "question": "Does `extracted_value` match the `field` as it appears in `source_text`?"}}, "customer_name": {"type": "... | {
"source_text": "Invoice #4471 issued March 3, 2026 to Beaver Dam Logistics for $12,840.00, net 30."
} | {
"invoice_number_is_correct": {
"type": "noul",
"instructions": {
"field": {
"name": "invoice_number",
"type": "string",
"description": "The identifier printed on the invoice."
},
"extracted_value": "4471",
"question": "Does `extracted_value` match the `field` ... | null | 4 | false | jev-latest | choice | Advanced: structure | Structured instructions | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"instructions",
"name",
"invoice",
"source",
"text",
"beaver"
] | TypeSafe AI - https://typesafe.ai |
29 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'JSON rubric for boundary clarification'. | I ordered the standing desk two weeks ago and tracking still says label created. Was I even charged? | {"department": {"type": "choice", "instructions": {"question": "Which team should handle this message?", "focus": "Classify the customer's primary request, not every topic mentioned."}, "criteria": {"billing": {"what": "Charges, invoices, refunds, or subscriptions", "not_for": "Order tracking or account access", "examp... | "I ordered the standing desk two weeks ago and tracking still says label created. Was I even charged?" | {
"department": {
"type": "choice",
"instructions": {
"question": "Which team should handle this message?",
"focus": "Classify the customer's primary request, not every topic mentioned."
},
"criteria": {
"billing": {
"what": "Charges, invoices, refunds, or subscriptions",
... | null | 1 | false | jev-latest | choice | Advanced: structure | JSON rubric for boundary clarification | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"account",
"boundary",
"charges",
"examples",
"order",
"access"
] | TypeSafe AI - https://typesafe.ai |
30 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Walking a taxonomy'. | 32oz plastic bottle with a flip straw lid. Fits most bike cages. | {"department": {"type": "choice", "instructions": "Which top-level department does this product belong to?", "criteria": {"Sporting Goods": {"Cycling": ["Bike Bottles & Cages", "Bike Lights", "Helmets"], "Fitness": ["Yoga Mats", "Resistance Bands"], "Outdoor": ["Tents", "Sleeping Bags", "Hydration Packs"]}, "Home & Kit... | "32oz plastic bottle with a flip straw lid. Fits most bike cages." | {
"department": {
"type": "choice",
"instructions": "Which top-level department does this product belong to?",
"criteria": {
"Sporting Goods": {
"Cycling": [
"Bike Bottles & Cages",
"Bike Lights",
"Helmets"
],
"Fitness": [
"Yoga Mats",
... | null | 1 | false | jev-latest | choice | Advanced: structure | Walking a taxonomy | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"bike",
"department",
"bottles",
"cages",
"bottle",
"branch"
] | TypeSafe AI - https://typesafe.ai |
31 | INSTRUCT_JEV | Define the TypeSafe System One Score question(s) used for 'Structured Score levels'. | Fixed the null check in the payment handler. Also refactored the retry loop while I was in there, and bumped the SDK version since the old one had that timeout bug. | {"pr_scope": {"type": "score", "instructions": {"question": "How focused is this pull request description on a single change?", "note": "Judge the number of independent changes, not the size of any one change."}, "criteria": [{"summary": "One change, clearly stated", "signals": ["A single fix or feature", "Nothing desc... | "Fixed the null check in the payment handler. Also refactored the retry loop while I was in there, and bumped the SDK version since the old one had that timeout bug." | {
"pr_scope": {
"type": "score",
"instructions": {
"question": "How focused is this pull request description on a single change?",
"note": "Judge the number of independent changes, not the size of any one change."
},
"criteria": [
{
"summary": "One change, clearly stated",
... | null | 1 | false | jev-latest | score | Advanced: structure | Structured Score levels | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"score",
"change",
"changes",
"signals",
"summary",
"also",
"criteria"
] | TypeSafe AI - https://typesafe.ai |
32 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Structured Noul criteria'. | {"sender": {"display_name": "Beaver Dam Builders Ltd.", "email": "donotreply@payroll.example"}, "message": "Your Q3 bonus is ready. Reply with your login password so we can verify your identity and release the funds."} | {"requests_credentials": {"type": "noul", "instructions": {"question": "Does the `message` ask the recipient to disclose a sensitive credential?", "inspect": "message", "focus": "Look for a request to send the credential itself, not a request to change or reset it."}, "criteria": {"true": {"what": "Asks the recipient t... | {
"sender": {
"display_name": "Beaver Dam Builders Ltd.",
"email": "donotreply@payroll.example"
},
"message": "Your Q3 bonus is ready. Reply with your login password so we can verify your identity and release the funds."
} | {
"requests_credentials": {
"type": "noul",
"instructions": {
"question": "Does the `message` ask the recipient to disclose a sensitive credential?",
"inspect": "message",
"focus": "Look for a request to send the credential itself, not a request to change or reset it."
},
"criteria":... | null | 1 | false | jev-latest | noul | Advanced: structure | Structured Noul criteria | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"noul",
"criteria",
"password",
"credential",
"examples",
"message",
"reply"
] | TypeSafe AI - https://typesafe.ai |
33 | INSTRUCT_JEV | Explain 'Advanced: structure' from the TypeSafe AI Choice documentation for System One (Jev). | Instructions, Choice options, Score levels, and Noul criteria all accept JSON structure. | # Advanced: structure
> Instructions, Choice options, Score levels, and Noul criteria all accept JSON structure.
System One models are trained to understand structure. | null | null | null | 0 | false | jev-latest | choice | Advanced: structure | Advanced: structure | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"structure",
"advanced",
"accept",
"criteria",
"instructions",
"json"
] | TypeSafe AI - https://typesafe.ai |
34 | INSTRUCT_JEV | Explain 'Where structure is allowed' from the TypeSafe AI Choice documentation for System One (Jev). | Instructions, Choice options, Score levels, and Noul criteria all accept JSON structure. | ## Where structure is allowed
Every one of these fields is an [`EntryType`](/sdk/javascript/api/type-aliases/EntryType).
| Field | Applies to | Accepted shape |
| --------------------------------------- | ------------------- | ------------------------... | null | null | null | 0 | false | jev-latest | choice | Advanced: structure | Where structure is allowed | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"criteria",
"allowed",
"descriptions",
"entrytype",
"noul",
"score"
] | TypeSafe AI - https://typesafe.ai |
35 | INSTRUCT_JEV | Explain 'When to structure a question' from the TypeSafe AI Noul documentation for System One (Jev). | Instructions, Choice options, Score levels, and Noul criteria all accept JSON structure. | ## When to structure a question
* **When it helps with clarity.** When a question has multiple parts, putting them in the form of JSON helps with clarity because the keys are labeled.
* **When question needs supporting data.** A schema, a taxonomy, or a database row is already JSON. Use the JSON entirely or pass in th... | null | null | null | 0 | false | jev-latest | noul | Advanced: structure | When to structure a question | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"noul",
"json",
"clarity",
"helps",
"structure",
"already",
"because"
] | TypeSafe AI - https://typesafe.ai |
36 | INSTRUCT_JEV | Explain 'Structured Choice options' from the TypeSafe AI Choice documentation for System One (Jev). | Instructions, Choice options, Score levels, and Noul criteria all accept JSON structure. | ## Structured Choice options
A Choice option description can be a structured object as well. | null | null | null | 0 | false | jev-latest | choice | Advanced: structure | Structured Choice options | _primitives | _primitives/advance_structure.md | [
"typesafe",
"jev",
"system-one",
"choice",
"structured",
"options",
"description",
"option",
"well"
] | TypeSafe AI - https://typesafe.ai |
37 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Choice question(s) for 'Request structure'. | My running shoes arrived in the wrong size. Can I swap them for a size 10? | {"department": {"type": "choice", "choice": "returns", "confidence": 1.0, "probabilities": {"shipping": 0.0, "returns": 1.0, "billing": 0.0}}} | "My running shoes arrived in the wrong size. Can I swap them for a size 10?" | {
"department": {
"type": "choice",
"instructions": "Which team should handle this?",
"criteria": {
"returns": "Exchanges, refunds, wrong or damaged items",
"shipping": "Delivery status, delays, lost packages",
"billing": "Charges, invoices, payment problems"
}
}
} | {
"department": {
"type": "choice",
"choice": "returns",
"confidence": 1,
"probabilities": {
"shipping": 0,
"returns": 1,
"billing": 0
}
}
} | 1 | true | jev-latest | choice | Choice | Request structure | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"model",
"request",
"criteria",
"instructions",
"client",
"department"
] | TypeSafe AI - https://typesafe.ai |
38 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Choice question(s) for 'A more complex example'. | Shoes arrived two weeks late and in the wrong size. Also I see two charges on my card. What are you going to do about this? | {"department": {"type": "choice", "choice": "returns", "confidence": 0.39, "probabilities": {"shipping": 0.02, "billing": 0.38, "returns": 0.6}}, "return_reason": {"type": "choice", "choice": "wrong_size", "confidence": 1.0, "probabilities": {"wrong_size": 1.0, "wrong_item": 0.0, "other": 0.0, "changed_mind": 0.0, "dam... | "Shoes arrived two weeks late and in the wrong size. Also I see two charges on my card. What are you going to do about this?" | {
"department": {
"type": "choice",
"instructions": "Which team should handle this?",
"criteria": {
"returns": "Exchanges, refunds, wrong or damaged items",
"shipping": "Delivery status, delays, lost packages",
"billing": "Charges, invoices, payment problems"
}
},
"return_reason"... | {
"department": {
"type": "choice",
"choice": "returns",
"confidence": 0.39,
"probabilities": {
"shipping": 0.02,
"billing": 0.38,
"returns": 0.6000000000000001
}
},
"return_reason": {
"type": "choice",
"choice": "wrong_size",
"confidence": 1,
"probabilities":... | 5 | true | jev-latest | choice | Choice | A more complex example | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"shipping",
"ticket",
"wrong",
"customer",
"team",
"department"
] | TypeSafe AI - https://typesafe.ai |
39 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Choice question(s) for 'Structured instructions and criteria'. | I sent the shoes back a week ago. When do I get my money? | {"return_topic": {"type": "choice", "choice": "return_status", "confidence": 1.0, "probabilities": {"return_policy": 0.0, "return_status": 1.0}}} | "I sent the shoes back a week ago. When do I get my money?" | {
"return_topic": {
"type": "choice",
"instructions": {
"question": "Which returns topic is the customer asking about?",
"focus": "Classify the information the customer wants."
},
"criteria": {
"return_policy": {
"what": "Whether and how an item can be returned",
"not... | {
"return_topic": {
"type": "choice",
"choice": "return_status",
"confidence": 1,
"probabilities": {
"return_policy": 0,
"return_status": 1
}
}
} | 1 | true | jev-latest | choice | Choice | Structured instructions and criteria | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"return",
"option",
"status",
"names",
"criteria",
"examples"
] | TypeSafe AI - https://typesafe.ai |
40 | INSTRUCT_JEV | Explain 'Choice' from the TypeSafe AI Choice documentation for System One (Jev). | A Choice is a System One question type for selecting one option from a defined set. The answer includes the selected option, a probability for each option, and confidence. | # Choice
> A Choice is a System One question type for selecting one option from a defined set. The answer includes the selected option, a probability for each option, and confidence.
Use a Choice when the answer is one of a fixed set of options. For example, which team handles a ticket, which category a product belon... | null | null | null | 0 | false | jev-latest | choice | Choice | Choice | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"option",
"answer",
"options",
"primitives",
"selected",
"category"
] | TypeSafe AI - https://typesafe.ai |
41 | INSTRUCT_JEV | Explain 'Response structure' from the TypeSafe AI Choice documentation for System One (Jev). | A Choice is a System One question type for selecting one option from a defined set. The answer includes the selected option, a probability for each option, and confidence. | ## Response structure
The response has one entry in `answers` per question, under the ids from the request. This is the response to the example request above:
```json theme={null}
{
"model": "jev-latest",
"answers": {
"department": {
"type": "choice",
"choice": "returns",
"confidence": 1.0,
... | null | null | null | 0 | false | jev-latest | choice | Choice | Response structure | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"probability",
"response",
"returns",
"option",
"probabilities"
] | TypeSafe AI - https://typesafe.ai |
42 | INSTRUCT_JEV | Explain 'Good practice: ask more than one question per call' from the TypeSafe AI Choice documentation for System One (Jev). | A Choice is a System One question type for selecting one option from a defined set. The answer includes the selected option, a probability for each option, and confidence. | ## Good practice: ask more than one question per call
Ask every Choice question your code might need in a single request rather than one request per question. Questions are evaluated in parallel. Adding questions barely changes the response time, and the code can ignore answers it doesn't need. Extra questions still c... | null | null | null | 0 | false | jev-latest | choice | Choice | Good practice: ask more than one question per call | _primitives | _primitives/primitives_choice.md | [
"typesafe",
"jev",
"system-one",
"choice",
"than",
"call",
"level",
"options",
"single",
"adding"
] | TypeSafe AI - https://typesafe.ai |
43 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Noul question(s) for 'Request'. | I have asked three times now. Can I please just talk to a real person? | {"is_human_escalation": {"type": "noul", "noul": 0.99}, "is_repeat_contact": {"type": "noul", "noul": 0.93}} | "I have asked three times now. Can I please just talk to a real person?" | {
"is_human_escalation": {
"type": "noul",
"instructions": "Is the customer asking for a human agent?"
},
"is_repeat_contact": {
"type": "noul",
"instructions": "Has the customer contacted support about this before?",
"criteria": {
"true": "Mentions a prior attempt, ticket, or that they ... | {
"is_human_escalation": {
"type": "noul",
"noul": 0.99
},
"is_repeat_contact": {
"type": "noul",
"noul": 0.93
}
} | 2 | true | jev-latest | noul | Noul | Request | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"instructions",
"request",
"asked",
"before",
"contact",
"criteria"
] | TypeSafe AI - https://typesafe.ai |
44 | INSTRUCT_JEV | Explain 'Noul' from the TypeSafe AI Noul documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | # Noul
> A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes.
Use a Noul when the answer is yes or no. For example, does this message ask for a refund, does this resume mention distributed systems, does this comment contain personal data. If the answer is one... | null | null | null | 0 | false | jev-latest | noul | Noul | Noul | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"answer",
"primitives",
"choice",
"probability",
"score",
"asks"
] | TypeSafe AI - https://typesafe.ai |
45 | INSTRUCT_JEV | Explain 'Writing a Noul question' from the TypeSafe AI Noul documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | ## Writing a Noul question
A Noul question evaluates a single yes/no question (or statement). It is defined by its `instructions`: the yes/no question to evaluate. It's good practice to phrase it so a high probability means "yes", so that the returned answer is unambiguous in its meaning.
You can optionally add `crit... | null | null | null | 0 | false | jev-latest | noul | Noul | Writing a Noul question | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"criteria",
"means",
"writing",
"answer",
"better",
"clarify"
] | TypeSafe AI - https://typesafe.ai |
46 | INSTRUCT_JEV | Explain 'Response' from the TypeSafe AI Noul documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | ## Response
```json theme={null}
{
"model": "jev-latest",
"answers": {
"is_human_escalation": {
"type": "noul",
"noul": 0.99
},
"is_repeat_contact": {
"type": "noul",
"noul": 0.93
}
},
"usage": {
"input_tokens": 360,
"output_tokens": 39
}
}
```
`noul` ranges f... | null | null | null | 0 | false | jev-latest | noul | Noul | Response | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"response",
"tokens",
"answer",
"answers",
"boolean",
"code"
] | TypeSafe AI - https://typesafe.ai |
47 | INSTRUCT_JEV | Explain 'Noul does not return a separate confidence value' from the TypeSafe AI Score documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | ## Noul does not return a separate confidence value
A value near 1 means a strong yes. A value near 0 means a strong no. A value near 0.5 gives yes and no similar probability.
For "Is the candidate strong in Python?", define what "strong" means. An unclear definition makes the probability hard to interpret. A value o... | null | null | null | 0 | false | jev-latest | score | Noul | Noul does not return a separate confidence value | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"score",
"strong",
"means",
"near",
"confidence",
"noul",
"primitives"
] | TypeSafe AI - https://typesafe.ai |
48 | INSTRUCT_JEV | Explain 'Example questions' from the TypeSafe AI Noul documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | ## Example questions
```
"Is the customer requesting a refund?"
"Does this resume mention experience with distributed systems?"
"Does the message contain personally identifiable information?"
"Does the room have a minifridge?"
``` | null | null | null | 0 | false | jev-latest | noul | Noul | Example questions | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"example",
"contain",
"customer",
"distributed",
"experience",
"identifiable"
] | TypeSafe AI - https://typesafe.ai |
49 | INSTRUCT_JEV | Explain 'Tips and advanced usage' from the TypeSafe AI Noul documentation for System One (Jev). | A Noul question asks the model to evaluate a yes/no question and return the probability that the answer is yes. | ## Tips and advanced usage
* **Phrasing.** Beyond a plain question, you can phrase the instruction as a statement for the model to evaluate for truthfulness. For "the customer is requesting a refund", a value near 1 means the statement is true. Try both phrasings with your own data to see what works best.
* **Optional... | null | null | null | 0 | false | jev-latest | noul | Noul | Tips and advanced usage | _primitives | _primitives/primitives_noul.md | [
"typesafe",
"jev",
"system-one",
"noul",
"advanced",
"criteria",
"instruction",
"means",
"statement",
"tips"
] | TypeSafe AI - https://typesafe.ai |
50 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Score question(s) for 'Request structure'. | The export button crashes the settings page in Safari. It works in Chrome, but a few of our customers only use Safari. | {"bug_severity": {"type": "score", "score": 1.3, "confidence": 0.54, "legend": {"0": "Cosmetic; no impact to functionality", "1": "Broken or degraded feature, but workaround exists", "2": "Blocking issue; no workaround exists"}, "probabilities": {"0": 0.0, "1": 0.7, "2": 0.3}}} | "The export button crashes the settings page in Safari. It works in Chrome, but a few of our customers only use Safari." | {
"bug_severity": {
"type": "score",
"instructions": "How severe is the reported issue?",
"criteria": [
"Cosmetic; no impact to functionality",
"Broken or degraded feature, but workaround exists",
"Blocking issue; no workaround exists"
]
}
} | {
"bug_severity": {
"type": "score",
"score": 1.3,
"confidence": 0.54,
"legend": {
"0": "Cosmetic; no impact to functionality",
"1": "Broken or degraded feature, but workaround exists",
"2": "Blocking issue; no workaround exists"
},
"probabilities": {
"0": 0,
"1":... | 1 | true | jev-latest | score | Score | Request structure | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"request",
"model",
"state",
"criteria",
"exists",
"fields"
] | TypeSafe AI - https://typesafe.ai |
51 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Score question(s) for 'Splitting a complex judgment into several Score questions'. | Export to PDF fails with a spinner that never finishes. Some of our team say CSV export still works for them, others say it fails too. This is the third time I'm writing in and honestly I'm done. Steps: open any report, click Export, choose PDF. Chrome 128 on macOS. | {"severity": {"type": "score", "score": 1.24, "confidence": 0.63, "legend": {"0": "Cosmetic; no impact to functionality", "1": "Broken or degraded feature, but workaround exists", "2": "Blocking issue; no workaround exists"}, "probabilities": {"0": 0.0, "1": 0.76, "2": 0.24}}, "frustration": {"type": "score", "score": ... | "Export to PDF fails with a spinner that never finishes. Some of our team say CSV export still works for them, others say it fails too. This is the third time I'm writing in and honestly I'm done. Steps: open any report, click Export, choose PDF. Chrome 128 on macOS." | {
"severity": {
"type": "score",
"instructions": "How severe is the reported issue?",
"criteria": [
"Cosmetic; no impact to functionality",
"Broken or degraded feature, but workaround exists",
"Blocking issue; no workaround exists"
]
},
"frustration": {
"type": "score",
"... | {
"severity": {
"type": "score",
"score": 1.24,
"confidence": 0.63,
"legend": {
"0": "Cosmetic; no impact to functionality",
"1": "Broken or degraded feature, but workaround exists",
"2": "Blocking issue; no workaround exists"
},
"probabilities": {
"0": 0,
"1": 0.... | 3 | true | jev-latest | score | Score | Splitting a complex judgment into several Score questions | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"report",
"steps",
"severity",
"environment",
"frustration",
"answers"
] | TypeSafe AI - https://typesafe.ai |
52 | INSTRUCT_JEV | Given the state, define and answer the TypeSafe System One Score question(s) for 'Structured level descriptions'. | Export to PDF fails with a spinner that never finishes. Some of our team say CSV export still works for them, others say it fails too. | {"bug_severity": {"type": "score", "score": 1.06, "confidence": 0.91, "legend": {"0": {"what": "Cosmetic; no impact to functionality", "examples": ["typo in a label", "misaligned icon"]}, "1": {"what": "Broken or degraded feature, but workaround exists", "examples": ["export fails in one browser but works in another"]}... | "Export to PDF fails with a spinner that never finishes. Some of our team say CSV export still works for them, others say it fails too." | {
"bug_severity": {
"type": "score",
"instructions": "How severe is the reported issue?",
"criteria": [
{
"what": "Cosmetic; no impact to functionality",
"examples": [
"typo in a label",
"misaligned icon"
]
},
{
"what": "Broken or degra... | {
"bug_severity": {
"type": "score",
"score": 1.06,
"confidence": 0.91,
"legend": {
"0": {
"what": "Cosmetic; no impact to functionality",
"examples": [
"typo in a label",
"misaligned icon"
]
},
"1": {
"what": "Broken or degraded fe... | 1 | true | jev-latest | score | Score | Structured level descriptions | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"examples",
"level",
"example",
"fails",
"confidence",
"export"
] | TypeSafe AI - https://typesafe.ai |
53 | INSTRUCT_JEV | Explain 'Score' from the TypeSafe AI Score documentation for System One (Jev). | A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence. | # Score
> A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence.
Use a Score when the answer is a position on a spectrum you can describe in steps. For example, how severe a bug is, how happy a customer i... | null | null | null | 0 | false | jev-latest | score | Score | Score | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"answer",
"levels",
"experience",
"primitives",
"between",
"candidate"
] | TypeSafe AI - https://typesafe.ai |
54 | INSTRUCT_JEV | Explain 'Levels' from the TypeSafe AI Score documentation for System One (Jev). | A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence. | ### Levels
Each entry in `criteria` is a level: one point on the spectrum of possible answers, described in words. A level's number is its position in the `criteria` array, starting at 0, so the three entries above are levels 0, 1 and 2. The order of the array is the numbering.
The model gets the descriptions and not... | null | null | null | 0 | false | jev-latest | score | Score | Levels | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"level",
"levels",
"client",
"criteria",
"model",
"response"
] | TypeSafe AI - https://typesafe.ai |
55 | INSTRUCT_JEV | Explain 'Response structure' from the TypeSafe AI Score documentation for System One (Jev). | A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence. | ## Response structure
The response has one entry in `answers` per question, under the ids from the request. This is the response to the example request above:
```json theme={null}
{
"model": "jev-latest",
"answers": {
"bug_severity": {
"type": "score",
"score": 1.3,
"confidence": 0.54,
... | null | null | null | 0 | false | jev-latest | score | Score | Response structure | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"level",
"confidence",
"number",
"probabilities",
"workaround",
"legend"
] | TypeSafe AI - https://typesafe.ai |
56 | INSTRUCT_JEV | Explain 'Reading a Score' from the TypeSafe AI Score documentation for System One (Jev). | A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence. | ## Reading a Score
Let's look at how the score changes with different inputs. For example, using the question and its levels from the request above:
```
"How severe is the reported issue?"
→ 0: Cosmetic; no impact to functionality
→ 1: Broken or degraded feature, but workaround exists
→ 2: Blocking issue; no wo... | null | null | null | 0 | false | jev-latest | score | Score | Reading a Score | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"level",
"code",
"style",
"confidence",
"export",
"width"
] | TypeSafe AI - https://typesafe.ai |
57 | INSTRUCT_JEV | Explain 'Writing good levels' from the TypeSafe AI Score documentation for System One (Jev). | A Score is a System One question type for rating content against ordered, descriptive levels. The answer includes a score, a probability for each level, and confidence. | ## Writing good levels
Describe situations, not degrees. "Broken or degraded feature, but workaround exists" gives the model something to match the state against. "Moderately severe" doesn't. Concrete descriptions can help the model distinguish levels. Check the answers against known examples; higher confidence alone ... | null | null | null | 0 | false | jev-latest | score | Score | Writing good levels | _primitives | _primitives/primitives_score.md | [
"typesafe",
"jev",
"system-one",
"score",
"levels",
"level",
"against",
"confidence",
"model",
"describe"
] | TypeSafe AI - https://typesafe.ai |
58 | INSTRUCT_JEV | Explain 'AI primer' from the TypeSafe AI Noul documentation for System One (Jev). | Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text. | # AI primer
> Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text.
Most AI products are built around a conversation between a model and a person. TypeSafe starts from a different bet: large-scale automation will be dominated by AI-to-AI and AI-to-software interac... | null | null | null | 0 | false | jev-latest | noul | AI primer | AI primer | _typesafe_foundations | _typesafe_foundations/ai_primer.md | [
"typesafe",
"jev",
"system-one",
"noul",
"interface",
"machine",
"primer",
"ai-to-ai",
"ai-to-software",
"around"
] | TypeSafe AI - https://typesafe.ai |
59 | INSTRUCT_JEV | Explain 'Building prod, not God' from the TypeSafe AI Noul documentation for System One (Jev). | Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text. | ## Building prod, not God
TypeSafe is not trying to build a model that does everything. It is designed for production systems where code needs a narrow decision it can inspect and act on.
Our expectation is that large-scale AI automation will be closer to 99% machine-to-machine interactions and 1% human interaction. ... | null | null | null | 0 | false | jev-latest | noul | AI primer | Building prod, not God | _typesafe_foundations | _typesafe_foundations/ai_primer.md | [
"typesafe",
"jev",
"system-one",
"noul",
"building",
"manifesto",
"prod",
"read",
"automation",
"behave"
] | TypeSafe AI - https://typesafe.ai |
60 | INSTRUCT_JEV | Explain 'Three post-training approaches' from the TypeSafe AI Noul documentation for System One (Jev). | Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text. | ## Three post-training approaches
Pretrained language models have been adapted in two major ways. TypeSafe adds a third. RLHF and RLVR are shown here for context; TypeSafe's training path is RLCD.
<Columns cols={3}>
<Card title="RLHF" icon="messages-square" type="note">
**Reinforcement learning from human feedb... | null | null | null | 0 | false | jev-latest | noul | AI primer | Three post-training approaches | _typesafe_foundations | _typesafe_foundations/ai_primer.md | [
"typesafe",
"jev",
"system-one",
"noul",
"card",
"models",
"rlhf",
"afvnpmcix68npsv1",
"ai-primer",
"images"
] | TypeSafe AI - https://typesafe.ai |
61 | INSTRUCT_JEV | Explain 'RLCD and calibrated decisions' from the TypeSafe AI Noul documentation for System One (Jev). | Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text. | ## RLCD and calibrated decisions
RLCD optimizes for a different output contract:
* The model does not generate text.
* It returns decisions and probabilities.
* Higher probability should correspond to a greater chance that the answer is correct.
Calibration makes uncertainty usable by software. Across many predictio... | null | null | null | 0 | false | jev-latest | noul | AI primer | RLCD and calibrated decisions | _typesafe_foundations | _typesafe_foundations/ai_primer.md | [
"typesafe",
"jev",
"system-one",
"noul",
"should",
"probability",
"about",
"assigned",
"decisions",
"occur"
] | TypeSafe AI - https://typesafe.ai |
62 | INSTRUCT_JEV | Explain 'The problems with RLHF' from the TypeSafe AI Noul documentation for System One (Jev). | Why TypeSafe trains decision models with calibrated probabilities instead of optimizing for generated text. | ## The problems with RLHF
RLHF teaches a model to say things that people prefer. That objective works well for chatbots, but it can also reward sycophancy and confident-sounding hallucinations.
Preference optimization also causes **mode dropping**: the model learns to favor a particular style, such as instruction fol... | null | null | null | 0 | false | jev-latest | noul | AI primer | The problems with RLHF | _typesafe_foundations | _typesafe_foundations/ai_primer.md | [
"typesafe",
"jev",
"system-one",
"noul",
"afvnpmcix68npsv1",
"ai-primer",
"images",
"webp",
"mode",
"rlhf"
] | TypeSafe AI - https://typesafe.ai |
63 | INSTRUCT_JEV | Explain 'Example use cases' from the TypeSafe AI Noul documentation for System One (Jev). | Explore TypeSafe use cases by industry and turn promising ideas into software workflows. | # Example use cases
> Explore TypeSafe use cases by industry and turn promising ideas into software workflows.
Use this map to brainstorm where TypeSafe could fit in your industry. Open the closest industry, scan the example decisions, and adapt them to the documents and actions in your own workflow. | null | null | null | 0 | false | jev-latest | noul | Example use cases | Example use cases | _typesafe_foundations | _typesafe_foundations/example_use_case.md | [
"typesafe",
"jev",
"system-one",
"noul",
"cases",
"example",
"industry",
"actions",
"adapt",
"brainstorm"
] | TypeSafe AI - https://typesafe.ai |
64 | INSTRUCT_JEV | Explain 'Example use case categories' from the TypeSafe AI Choice documentation for System One (Jev). | Explore TypeSafe use cases by industry and turn promising ideas into software workflows. | ## Example use case categories
<Columns cols={2}>
<Card title="AI Automation Software" icon="blocks">
Interleave AI with reliable software in a way where you can run it a million times in the background without a human co-pilot. Code owns control flow (not markdown files) while TypeSafe handles the semantic deci... | null | null | null | 0 | false | jev-latest | choice | Example use cases | Example use case categories | _typesafe_foundations | _typesafe_foundations/example_use_case.md | [
"typesafe",
"jev",
"system-one",
"choice",
"card",
"icon",
"title",
"make",
"giant",
"other"
] | TypeSafe AI - https://typesafe.ai |
65 | INSTRUCT_JEV | Explain 'Example automation use cases' from the TypeSafe AI Choice documentation for System One (Jev). | Explore TypeSafe use cases by industry and turn promising ideas into software workflows. | ## Example automation use cases
<AccordionGroup>
<Accordion title="Search and retrieval" icon="search">
* Replace or supplement embeddings in RAG pipelines with semantic search, scoring, and ranking.
* Score query-to-candidate relevance.
* Rerank results with pairwise comparisons.
* Cross-encode quer... | null | null | null | 0 | false | jev-latest | choice | Example use cases | Example automation use cases | _typesafe_foundations | _typesafe_foundations/example_use_case.md | [
"typesafe",
"jev",
"system-one",
"choice",
"accordion",
"icon",
"title",
"detect",
"claims",
"classify"
] | TypeSafe AI - https://typesafe.ai |
66 | INSTRUCT_JEV | Explain 'Example task categories' from the TypeSafe AI Choice documentation for System One (Jev). | Explore TypeSafe use cases by industry and turn promising ideas into software workflows. | ## Example task categories
| Decision shape | Reach for it when | Examples |
| ------------------------------ | ---------------------------------------------------------- | --------------------------... | null | null | null | 0 | false | jev-latest | choice | Example use cases | Example task categories | _typesafe_foundations | _typesafe_foundations/example_use_case.md | [
"typesafe",
"jev",
"system-one",
"choice",
"candidate",
"need",
"quality",
"search",
"semantic",
"categories"
] | TypeSafe AI - https://typesafe.ai |
67 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"ticket_message": "My flight was cancelled. Can I get a refund?", "refund_policy": "Cancelled flights are eligible for a full refund."} | {"policy_supports_refund": {"type": "noul", "instructions": "Does the refund policy support the refund requested in the ticket?"}} | {
"ticket_message": "My flight was cancelled. Can I get a refund?",
"refund_policy": "Cancelled flights are eligible for a full refund."
} | {
"policy_supports_refund": {
"type": "noul",
"instructions": "Does the refund policy support the refund requested in the ticket?"
}
} | null | 1 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
68 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"support": {"tickets": [{"message": "I was charged twice for order A-104."}, {"message": "How do I reset my password?"}]}, "commerce": {"orders": [{"id": "A-104", "charges": [{"amount_usd": 49, "status": "captured"}, {"amount_usd": 49, "status": "captured"}]}]}, "account": {"security": {"password_reset": "Email a rese... | {"duplicate_charge": {"type": "noul", "instructions": "Do `support.tickets[0].message` and `commerce.orders[0].charges` indicate a duplicate charge?"}, "password_reset_supported": {"type": "noul", "instructions": "Can `account.security.password_reset` resolve the request in `support.tickets[1].message`?"}} | {
"support": {
"tickets": [
{
"message": "I was charged twice for order A-104."
},
{
"message": "How do I reset my password?"
}
]
},
"commerce": {
"orders": [
{
"id": "A-104",
"charges": [
{
"amount_usd": 49,
... | {
"duplicate_charge": {
"type": "noul",
"instructions": "Do `support.tickets[0].message` and `commerce.orders[0].charges` indicate a duplicate charge?"
},
"password_reset_supported": {
"type": "noul",
"instructions": "Can `account.security.password_reset` resolve the request in `support.tickets[1]... | null | 2 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
69 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"message": {"sender": {"display_name": "Acme Payroll", "email": "rewards@claim-bonus.example"}, "subject": "Urgent: claim your employee bonus", "body": "You have been selected for a $1,000 bonus. Confirm your payroll password today to receive it.", "links": [{"text": "Claim bonus", "url": "http://claim-bonus.example/a... | {"is_spam": {"type": "noul", "instructions": "Is `message` spam?"}} | {
"message": {
"sender": {
"display_name": "Acme Payroll",
"email": "rewards@claim-bonus.example"
},
"subject": "Urgent: claim your employee bonus",
"body": "You have been selected for a $1,000 bonus. Confirm your payroll password today to receive it.",
"links": [
{
"text... | {
"is_spam": {
"type": "noul",
"instructions": "Is `message` spam?"
}
} | null | 1 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
70 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"message": {"sender": {"display_name": "Acme Payroll", "email": "rewards@claim-bonus.example"}, "subject": "Urgent: claim your employee bonus", "body": "You have been selected for a $1,000 bonus. Confirm your payroll password today to receive it.", "links": [{"text": "Claim bonus", "url": "http://claim-bonus.example/a... | {"requests_credentials": {"type": "noul", "instructions": "Does `message.body` ask the recipient to provide a password or other login credential?"}, "offers_unexpected_reward": {"type": "noul", "instructions": "Does `message.body` claim the recipient received an unexpected prize, payment, or reward?"}, "creates_time_pr... | {
"message": {
"sender": {
"display_name": "Acme Payroll",
"email": "rewards@claim-bonus.example"
},
"subject": "Urgent: claim your employee bonus",
"body": "You have been selected for a $1,000 bonus. Confirm your payroll password today to receive it.",
"links": [
{
"text... | {
"requests_credentials": {
"type": "noul",
"instructions": "Does `message.body` ask the recipient to provide a password or other login credential?"
},
"offers_unexpected_reward": {
"type": "noul",
"instructions": "Does `message.body` claim the recipient received an unexpected prize, payment, or r... | null | 6 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
71 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"request": {"text": "What's the weather in Seattle tomorrow in Fahrenheit?", "location": "Seattle, WA", "date": "2026-09-03", "unit": "fahrenheit"}, "available_tools": {"geocode_city": {"description": "Resolve a city to latitude and longitude.", "parameters": {"city": "string"}}, "get_weather": {"description": "Get th... | {"tool_calls_are_correct": {"type": "noul", "instructions": "Is `trace.tool_calls` correct for `request` and `available_tools`?"}} | {
"request": {
"text": "What's the weather in Seattle tomorrow in Fahrenheit?",
"location": "Seattle, WA",
"date": "2026-09-03",
"unit": "fahrenheit"
},
"available_tools": {
"geocode_city": {
"description": "Resolve a city to latitude and longitude.",
"parameters": {
"city"... | {
"tool_calls_are_correct": {
"type": "noul",
"instructions": "Is `trace.tool_calls` correct for `request` and `available_tools`?"
}
} | null | 1 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
72 | INSTRUCT_JEV | Define the TypeSafe System One Noul question(s) used for 'Design a System One workflow'. | {"request": {"text": "What's the weather in Seattle tomorrow in Fahrenheit?", "location": "Seattle, WA", "date": "2026-09-03", "unit": "fahrenheit"}, "available_tools": {"geocode_city": {"description": "Resolve a city to latitude and longitude.", "parameters": {"city": "string"}}, "get_weather": {"description": "Get th... | {"geocode_tool_is_relevant": {"type": "noul", "instructions": "Is `trace.tool_calls[0].name` an appropriate tool for resolving `request.location`?"}, "geocode_location_matches": {"type": "noul", "instructions": "Does `trace.tool_calls[0].arguments.city` match `request.location`?"}, "geocode_arguments_match_schema": {"t... | {
"request": {
"text": "What's the weather in Seattle tomorrow in Fahrenheit?",
"location": "Seattle, WA",
"date": "2026-09-03",
"unit": "fahrenheit"
},
"available_tools": {
"geocode_city": {
"description": "Resolve a city to latitude and longitude.",
"parameters": {
"city"... | {
"geocode_tool_is_relevant": {
"type": "noul",
"instructions": "Is `trace.tool_calls[0].name` an appropriate tool for resolving `request.location`?"
},
"geocode_location_matches": {
"type": "noul",
"instructions": "Does `trace.tool_calls[0].arguments.city` match `request.location`?"
},
"geoco... | null | 9 | false | jev-latest | noul | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"tool",
"instructions",
"title",
"example",
"message",
"accordion"
] | TypeSafe AI - https://typesafe.ai |
73 | INSTRUCT_JEV | Define the TypeSafe System One Choice question(s) used for 'Design a System One workflow'. | How many disposable virtual cards can I make per day? | {"card_help_topic": {"type": "choice", "instructions": {"question": "Which disposable virtual card topic is the user asking about?", "focus": "Classify the information the user wants."}, "criteria": {"get_disposable_virtual_card": {"what": "Purpose, eligibility, or setup", "not_for": "Quantity, transaction, or merchant... | "How many disposable virtual cards can I make per day?" | {
"card_help_topic": {
"type": "choice",
"instructions": {
"question": "Which disposable virtual card topic is the user asking about?",
"focus": "Classify the information the user wants."
},
"criteria": {
"get_disposable_virtual_card": {
"what": "Purpose, eligibility, or setu... | null | 1 | false | jev-latest | choice | How to build with TypeSafe | Design a System One workflow | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"choice",
"noul",
"tool",
"instructions",
"title",
"example",
"message"
] | TypeSafe AI - https://typesafe.ai |
74 | INSTRUCT_JEV | Explain 'How to build with TypeSafe' from the TypeSafe AI Noul documentation for System One (Jev). | Design AI-powered software by keeping code in control and giving System One narrow, structured decisions. | # How to build with TypeSafe
> Design AI-powered software by keeping code in control and giving System One narrow, structured decisions.
System One is TypeSafe's model for building AI-powered software, not agents. It does not generate code or choose its own next action. It provides AI primitives that embed into softw... | null | null | null | 0 | false | jev-latest | noul | How to build with TypeSafe | How to build with TypeSafe | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"code",
"software",
"build",
"control",
"system",
"ai-powered"
] | TypeSafe AI - https://typesafe.ai |
75 | INSTRUCT_JEV | Explain 'Three software architectures' from the TypeSafe AI Score documentation for System One (Jev). | Design AI-powered software by keeping code in control and giving System One narrow, structured decisions. | ## Three software architectures
TypeSafe is designed for building **AI-powered software**, where code owns the workflow and AI handles narrow, structured decisions.
<Tabs>
<Tab title="Traditional software">
Traditional code is a complex decision tree made from simple software primitives. Because each primitive ... | null | null | null | 0 | false | jev-latest | score | How to build with TypeSafe | Three software architectures | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"score",
"software",
"afvnpmcix68npsv1",
"ai-powered",
"architectures",
"how-to-build-with-typesafe",
"images"
] | TypeSafe AI - https://typesafe.ai |
76 | INSTRUCT_JEV | Explain 'What makes System One composable' from the TypeSafe AI Noul documentation for System One (Jev). | Design AI-powered software by keeping code in control and giving System One narrow, structured decisions. | ## What makes System One composable
<Columns cols={2}>
<Card title="Structured" icon="braces">
System One is type-safe by construction. Decisions and probabilities conform to the structured software types and JSON schema your code expects, so it never has to recover a value from generated prose.
</Card>
<Ca... | null | null | null | 0 | false | jev-latest | noul | How to build with TypeSafe | What makes System One composable | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"card",
"icon",
"title",
"system",
"calibrated",
"columns"
] | TypeSafe AI - https://typesafe.ai |
77 | INSTRUCT_JEV | Explain 'Putting it all together' from the TypeSafe AI Noul documentation for System One (Jev). | Design AI-powered software by keeping code in control and giving System One narrow, structured decisions. | ## Putting it all together
This support-ticket workflow keeps deterministic work in code, sends only relevant structured context, evaluates many atomic questions in one request, and composes the answers with explicit confidence gates.
```python title="triage_ticket.py" theme={null}
from typesafe_sdk import Choice, No... | null | null | null | 0 | false | jev-latest | noul | How to build with TypeSafe | Putting it all together | _typesafe_foundations | _typesafe_foundations/how_to_build_with_typesafe.md | [
"typesafe",
"jev",
"system-one",
"noul",
"ticket",
"order",
"answers",
"examples",
"message",
"open"
] | TypeSafe AI - https://typesafe.ai |
78 | INSTRUCT_JEV | Explain 'Confidence' from the TypeSafe AI Choice documentation for System One (Jev). | How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior. | # Confidence
> How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior.
All Score and Choice answers from TypeSafe include a `probabilities` property representing the probability distribution across the options (for Choice) or levels (for Score). The *shape* of th... | null | null | null | 0 | false | jev-latest | choice | Confidence | Confidence | _typesafe_foundations | _typesafe_foundations/typesafe_foundations_confidence.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"answer",
"answers",
"distribution",
"means",
"probability"
] | TypeSafe AI - https://typesafe.ai |
79 | INSTRUCT_JEV | Explain 'Confidence is derived from the probabilities' from the TypeSafe AI Choice documentation for System One (Jev). | How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior. | ## Confidence is derived from the probabilities
`confidence` is a statistic computed from the probability distribution the answer already gives you. TypeSafe computes it for you and returns it on every Choice and Score answer, so the common case needs no extra work on your side.
<Note>
**A solid default:** We provi... | null | null | null | 0 | false | jev-latest | choice | Confidence | Confidence is derived from the probabilities | _typesafe_foundations | _typesafe_foundations/typesafe_foundations_confidence.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"distribution",
"probabilities",
"score",
"means",
"across"
] | TypeSafe AI - https://typesafe.ai |
80 | INSTRUCT_JEV | Explain '"I don't know" is a useful signal' from the TypeSafe AI Score documentation for System One (Jev). | How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior. | ## "I don't know" is a useful signal
If an intelligent system, whether human or machine, cannot express honest uncertainty, the system cannot be trusted.
Confidence gives you a built-in mechanism for the model to say "I'm not sure about this one." This lets your code implement different behavior for different levels ... | null | null | null | 0 | false | jev-latest | score | Confidence | "I don't know" is a useful signal | _typesafe_foundations | _typesafe_foundations/typesafe_foundations_confidence.md | [
"typesafe",
"jev",
"system-one",
"score",
"cannot",
"different",
"know",
"signal",
"system",
"useful"
] | TypeSafe AI - https://typesafe.ai |
81 | INSTRUCT_JEV | Explain 'Three paths for using confidence in your code' from the TypeSafe AI Choice documentation for System One (Jev). | How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior. | ## Three paths for using confidence in your code
A useful starting pattern is to divide confidence into three ranges, each producing a different system behavior:
**High confidence:** Act automatically. The model has a clear read and you can proceed without human involvement.
**Medium confidence:** Proceed with cauti... | null | null | null | 0 | false | jev-latest | choice | Confidence | Three paths for using confidence in your code | _typesafe_foundations | _typesafe_foundations/typesafe_foundations_confidence.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"model",
"three",
"code",
"different",
"human"
] | TypeSafe AI - https://typesafe.ai |
82 | INSTRUCT_JEV | Explain 'Thresholds scale with risk' from the TypeSafe AI Choice documentation for System One (Jev). | How TypeSafe reports certainty, how it differs from probability, and how to use it to control system behavior. | ## Thresholds scale with risk
A confidence threshold is not one number. Different actions within the same system should be gated at different levels depending on the consequences of getting it wrong.
```python theme={null}
response = client.system_one(
state=user_message,
questions={
"action": Choice(... | null | null | null | 0 | false | jev-latest | choice | Confidence | Thresholds scale with risk | _typesafe_foundations | _typesafe_foundations/typesafe_foundations_confidence.md | [
"typesafe",
"jev",
"system-one",
"choice",
"confidence",
"action",
"account",
"balance",
"user",
"approve"
] | TypeSafe AI - https://typesafe.ai |
83 | INSTRUCT_JEV | Summarize Build with TypeSafe from the TypeSafe AI System One documentation. | Build with TypeSafe | ---
name: typesafe-ai
license: MIT
description: >
Build AI-powered software with TypeSafe: small units of AI intelligence you
can use like programming primitives. Its System One models, including Jev,
turn natural language and application state into typed judgments and
probabilities that code can combine. Use w... | null | null | null | 0 | false | jev-latest | score | Build with TypeSafe | Build with TypeSafe | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"score",
"could",
"ai-powered",
"application",
"applications",
"become",
"brainstorming"
] | TypeSafe AI - https://typesafe.ai |
84 | INSTRUCT_JEV | Explain 'Build with TypeSafe' from the TypeSafe AI Noul documentation for System One (Jev). | Build with TypeSafe | # Build with TypeSafe
TypeSafe makes units of AI intelligence usable like programming primitives: small
judgments you can compose into larger capabilities. Its **System One models** return
fast, focused judgments that software can consume directly. **Jev** is TypeSafe's
flagship and first System One model. It understa... | null | null | null | 0 | false | jev-latest | noul | Build with TypeSafe | Build with TypeSafe | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"noul",
"build",
"code",
"judgments",
"model",
"system",
"answers"
] | TypeSafe AI - https://typesafe.ai |
85 | INSTRUCT_JEV | Explain 'Read the live docs' from the TypeSafe AI Choice documentation for System One (Jev). | Build with TypeSafe | ## Read the live docs
**The live TypeSafe docs are the source of truth. Read them as part of the task.**
This skill gives direction; the docs carry current concepts, prompting guidance,
API contracts, SDK usage, models, limits, and worked examples.
- Start with the [documentation index](https://docs.typesafe.ai/llms.... | null | null | null | 0 | false | jev-latest | choice | Build with TypeSafe | Read the live docs | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"choice",
"concepts",
"index",
"links",
"live",
"read",
"relevant"
] | TypeSafe AI - https://typesafe.ai |
86 | INSTRUCT_JEV | Explain 'Find the useful shape' from the TypeSafe AI Choice documentation for System One (Jev). | Build with TypeSafe | ## Find the useful shape
Start from the behavior the user wants: what will the application show, select,
change, or hand off? Work backward to the judgments it needs. Keep known rules,
calculations, exact lookups, and execution in code. Preserve the user's chosen stack
and scope; add TypeSafe where semantic understand... | null | null | null | 0 | false | jev-latest | choice | Build with TypeSafe | Find the useful shape | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"choice",
"cookbooks",
"code",
"explore",
"select",
"user",
"find"
] | TypeSafe AI - https://typesafe.ai |
87 | INSTRUCT_JEV | Explain 'Design the judgments' from the TypeSafe AI Score documentation for System One (Jev). | Build with TypeSafe | ## Design the judgments
Choose by what the answer means, then read the relevant primitive page:
| Need | Primitive | Important distinction |
| --- | --- | --- |
| One of a defined set | [Choice](https://docs.typesafe.ai/primitives/choice.md) | Picks one option; its distribution compares competing options |
| Whether ... | null | null | null | 0 | false | jev-latest | score | Build with TypeSafe | Design the judgments | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"score",
"judgment",
"primitives",
"answer",
"answers",
"choice",
"choose"
] | TypeSafe AI - https://typesafe.ai |
88 | INSTRUCT_JEV | Explain 'Compose and verify' from the TypeSafe AI Choice documentation for System One (Jev). | Build with TypeSafe | ## Compose and verify
**Ask independent questions over the same state together**, including useful
speculative questions. They run in parallel and cannot see one another's answers.
State each speculative premise explicitly; code consumes the applicable answers.
A second request is warranted when an earlier answer is n... | null | null | null | 0 | false | jev-latest | choice | Build with TypeSafe | Compose and verify | root | agents_prompts_jev.md.txt | [
"typesafe",
"jev",
"system-one",
"choice",
"state",
"answers",
"confidence",
"evidence",
"behavior",
"code"
] | TypeSafe AI - https://typesafe.ai |
89 | INSTRUCT_JEV | Explain 'System One' from the TypeSafe AI Noul documentation for System One (Jev). | System One models make fast, structured decisions for software. Jev is TypeSafe's flagship model and the first System One model. | # System One
> System One models make fast, structured decisions for software. Jev is TypeSafe's flagship model and the first System One model.
System One models are a class of AI models built to make fast, structured decisions that software can use directly. A System One model evaluates a [state](/concepts/state) an... | null | null | null | 0 | false | jev-latest | noul | System One | System One | root | docs_concepts_system_one.md | [
"typesafe",
"jev",
"system-one",
"noul",
"system",
"model",
"decisions",
"models",
"text",
"evaluates"
] | TypeSafe AI - https://typesafe.ai |
90 | INSTRUCT_JEV | Explain 'How it differs from an LLM' from the TypeSafe AI Noul documentation for System One (Jev). | System One models make fast, structured decisions for software. Jev is TypeSafe's flagship model and the first System One model. | ## How it differs from an LLM
System One models are trained for calibrated decisions: their probabilities are optimized against outcomes to reflect uncertainty. Calibration is measured across groups of predictions; it does not guarantee that an individual answer is correct.
System One models do not write replies, pro... | null | null | null | 0 | false | jev-latest | noul | System One | How it differs from an LLM | root | docs_concepts_system_one.md | [
"typesafe",
"jev",
"system-one",
"noul",
"system",
"primitives",
"choice",
"fast",
"frustrated",
"models"
] | TypeSafe AI - https://typesafe.ai |
91 | INSTRUCT_JEV | Explain 'Fast judgments inside a larger workflow' from the TypeSafe AI Choice documentation for System One (Jev). | System One models make fast, structured decisions for software. Jev is TypeSafe's flagship model and the first System One model. | ## Fast judgments inside a larger workflow
For a refund request, your application can:
1. Build a state containing the customer's message, the relevant transactions, and the refund policy.
2. Ask independent questions together: whether a refund was requested, whether the evidence indicates a duplicate charge, and whe... | null | null | null | 0 | false | jev-latest | choice | System One | Fast judgments inside a larger workflow | root | docs_concepts_system_one.md | [
"typesafe",
"jev",
"system-one",
"choice",
"refund",
"answers",
"combine",
"larger",
"system",
"whether"
] | TypeSafe AI - https://typesafe.ai |
92 | INSTRUCT_JEV | Explain 'Call a System One model' from the TypeSafe AI Noul documentation for System One (Jev). | System One models make fast, structured decisions for software. Jev is TypeSafe's flagship model and the first System One model. | ## Call a System One model
Call a System One model through one of our [client SDKs](/sdk) or `POST /v1/systemone` in the [HTTP API](/api). The `model` field selects which model handles the request. The examples in these docs use `jev-latest`, which is also the SDK default. See [Models](/models) for the available model... | null | null | null | 0 | false | jev-latest | noul | System One | Call a System One model | root | docs_concepts_system_one.md | [
"typesafe",
"jev",
"system-one",
"noul",
"model",
"call",
"models",
"system",
"primitives",
"state"
] | TypeSafe AI - https://typesafe.ai |
93 | INSTRUCT_JEV | Explain 'Introduction' from the TypeSafe AI Choice documentation for System One (Jev). | Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions; get structured answers your code can use directly. | # Introduction
> Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions; get structured answers your code can use directly.
Large language models (LLMs) are designed to produce text for humans to read. When you need a model to make a judgment that your code will consume, that ... | null | null | null | 0 | false | jev-latest | choice | Introduction | Introduction | root | introduction_.md | [
"typesafe",
"jev",
"system-one",
"choice",
"model",
"code",
"structured",
"system",
"directly",
"typed"
] | TypeSafe AI - https://typesafe.ai |
94 | INSTRUCT_JEV | Explain 'TypeSafe primitives' from the TypeSafe AI Score documentation for System One (Jev). | Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions; get structured answers your code can use directly. | ## TypeSafe primitives
TypeSafe exposes three *AI primitives*. Similar to software primitives, our AI primitives are modular, composable, structured, reliable, and fast. Each asks a different type of *question* and returns a different type of answer.
| Question type | Goal | Ret... | null | null | null | 0 | false | jev-latest | score | Introduction | TypeSafe primitives | root | introduction_.md | [
"typesafe",
"jev",
"system-one",
"score",
"primitives",
"choice",
"noul",
"adding",
"confidence",
"different"
] | TypeSafe AI - https://typesafe.ai |
95 | INSTRUCT_JEV | Explain 'Atomic questions, composed in code' from the TypeSafe AI Score documentation for System One (Jev). | Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions; get structured answers your code can use directly. | ## Atomic questions, composed in code
System One models work best when each question asks one specific, well-scoped thing. Think of each question as a gut-check determination: the kind of judgment a highly knowledgeable person could make in a few seconds given the right context.
If the question you want to ask would ... | null | null | null | 0 | false | jev-latest | score | Introduction | Atomic questions, composed in code | root | introduction_.md | [
"typesafe",
"jev",
"system-one",
"score",
"code",
"atomic",
"combine",
"composed",
"about",
"asks"
] | TypeSafe AI - https://typesafe.ai |
96 | INSTRUCT_JEV | Explain 'Next steps' from the TypeSafe AI Choice documentation for System One (Jev). | Jev is TypeSafe's flagship model and the first System One model. Send state and typed questions; get structured answers your code can use directly. | ## Next steps
* [Quick Start](/introduction/quickstart) — Everything you need to get started immediately.
* [AI Primer](/introduction/machine-learning-primer) — Why TypeSafe trains models for calibrated decisions instead of generated text.
* [Primitives (Questions)](/primitives) — How to define questions, choose betwe... | null | null | null | 0 | false | jev-latest | choice | Introduction | Next steps | root | introduction_.md | [
"typesafe",
"jev",
"system-one",
"choice",
"patterns",
"confidence",
"introduction",
"next",
"primitives",
"steps"
] | TypeSafe AI - https://typesafe.ai |
End of preview. Expand in Data Studio
INSTRUCT_JEV
INSTRUCT_JEV is an instruction corpus built from the TypeSafe AI documentation
for Jev, the first System One model. It is structured around the three TypeSafe
question primitives - Choice, Noul and Score - and mirrors the raw corpus
captured in deckerGUI-jev_corpus_RAW.
Credit
All documentation, concepts, API design and examples are the work of TypeSafe AI (https://typesafe.ai). Jev, System One and the Choice / Noul / Score primitives are TypeSafe's. This corpus is compiled and redistributed under their MIT-licensed public documentation at https://docs.typesafe.ai.
Corpus and concepts are the property of TypeSafe AI. Jev, System One, and the Choice / Noul / Score primitives are TypeSafe's. Used and redistributed under their MIT-licensed public documentation. Compiled by DeckerGUI. Source: TypeSafe AI documentation (Jev / System One) - https://docs.typesafe.ai Compiled by DeckerGUI.
Scope
The RAW mirror captures the full cleaned TypeSafe AI documentation corpus
(deckerGUI-jev_corpus_RAW, 128 files). The published instruction
rows are curated from it: every extracted typed Choice / Noul / Score example is kept,
plus the conceptual documentation (_primitives, _typesafe_foundations, _patterns,
_demos and the root pages). The large per-heading SDK / cookbook chunk set is kept in
the RAW mirror but is not expanded into instruction rows.
Function types
Every row is tagged with one of the three TypeSafe System One function types:
function_type |
Primitive | Answer shape |
|---|---|---|
choice |
Choice - pick one option from a list | choice, probabilities, confidence |
noul |
Noul - is the statement true? | noul (0-1) |
score |
Score - rate along ordered levels | score, legend, probabilities, confidence |
Stats
- Rows: 119
- choice: 47 / noul: 51 / score: 21
- Rows with an extracted typed question block: 24 (of which 7 include typed answers)
- Unique source documents: 114 (from 128 captured files)
Schema
| Field | Type | Description |
|---|---|---|
id |
int | Sequential row id |
dataset |
string | INSTRUCT_JEV |
instruction |
string | The task / question |
input |
string | Context or state |
output |
string | The expected answer |
state |
json | TypeSafe state when available |
question_block |
json | Typed Choice / Noul / Score question objects |
answer_block |
json | Typed answers |
question_count |
int | Number of questions in the block |
has_answer |
bool | Whether a typed answer block was extracted |
model |
string | Model alias shown in docs (jev-latest) |
function_type |
string | choice | noul | score |
doc_title |
string | Source document title |
section |
string | Source section heading |
category |
string | Corpus category |
source |
string | Raw corpus relative path |
tags |
array | Search tags |
credit |
string | Attribution |
Usage
from datasets import load_dataset
dataset = load_dataset("ctaxnagomi/INSTRUCT_JEV")
print(dataset["train"][0])
Per-primitive subsets are published alongside the full set: choice.jsonl,
noul.jsonl, score.jsonl.
License
MIT - see the upstream TypeSafe AI documentation for the original terms. Compiled by DeckerGUI.
- Downloads last month
- 12