Karmane/npm-ai-agent-mcp-packages-enriched

npm AI Agent and MCP Packages Enriched Dataset

This dataset packages public npm registry and download records for AI-agent, MCP, orchestration, and model-tooling packages into one analysis-ready dataframe.

Each row represents a package discovered through overlapping npm search terms and enriched with registry metadata, publishing recency, maintainer counts, dependency surface, CLI and TypeScript signals, provider mentions, commercialization cues, and official npm download metrics. The result is easier to use for developer-tool market research, ecosystem mapping, dependency discovery, package screening, and AI tooling analysis than manually unioning search results with package and download endpoints by hand.

---

Dataset Preview

record_id	package_name	package_scope	unscoped_name	package_description	latest_version	license_name	author_name	publisher_name	maintainer_count	published_at	first_published_at	package_age_days	days_since_publish	is_recently_published_90d	is_deprecated	deprecated_message	weekly_downloads	monthly_downloads	search_score	matched_search_terms	matched_search_term_count	primary_search_term	best_search_rank	mean_search_rank	keywords	keyword_count	dependency_count	dev_dependency_count	peer_dependency_count	bin_command_count	has_cli_bin	has_typescript_types	has_exports	engines_node	repository_url	repository_domain	homepage_url	homepage_domain	bugs_url	npm_package_url	registry_url	source_search_url	ecosystem_segment	provider_mentions	provider_count	has_core_ai_signal	has_mcp_signal	has_agent_signal	has_sdk_signal	has_commercial_terms	commercial_readiness	last_collected_at	monthly_download_rank	weekly_download_rank	downloads_tier
npm-package-openai	openai		openai	The official TypeScript library for the OpenAI API	6.39.1	Apache-2.0	OpenAI	GitHub Actions	17	2026-05-28T17:46:52.775Z	2020-07-09T13:31:41.309Z	2152	3	True	False		23430731	92583631	2360.22	openai	1	openai	1	1		0	0	0	2	1	True	True	True		https://github.com/openai/openai-node.git	github.com	https://github.com/openai/openai-node#readme	github.com	https://github.com/openai/openai-node/issues	https://www.npmjs.com/package/openai	https://registry.npmjs.org/openai	https://www.npmjs.com/search?q=openai	Developer Tooling / CLI	OpenAI	1	False	False	False	False	False	high	2026-06-01T02:14:11Z	1	1	1m+
npm-package-ai	ai		ai	AI SDK by Vercel - build apps like ChatGPT, Claude, Gemini, and more with a single interface for any... [106 more chars]	6.0.193	Apache-2.0		GitHub Actions	5	2026-05-28T23:37:21.208Z	2014-02-21T22:09:35.189Z	4482	3	True	False		12944419	56060745	224.889	anthropic	openai	2	anthropic	33	56.5	ai	vercel	sdk	llm	mcp	tool-calling	tools	structured-output	agent	agentic	generative	genai	chatbot	pr... [79 more chars]	23	4	10	1	0	False	True	True	>=18	https://github.com/vercel/ai.git	github.com	https://ai-sdk.dev/docs	ai-sdk.dev	https://github.com/vercel/ai/issues	https://www.npmjs.com/package/ai	https://registry.npmjs.org/ai
npm-package-@anthropic-ai/claude-code-linux-x64	@anthropic-ai/claude-code-linux-x64	@anthropic-ai	claude-code-linux-x64	Native binary for Claude Code on linux-x64	2.1.159	SEE LICENSE IN LICENSE.md		wolffiex	14	2026-05-31T17:49:20.818Z	2026-04-13T23:59:32.601Z	48	0	True	False		9308309	36162771	312.338	anthropic	claude	2	anthropic	2	41.5		0	0	0	0	0	False	False	False							https://www.npmjs.com/package/@anthropic-ai/claude-code-linux-x64	https://registry.npmjs.org/%40anthropic-ai%2Fclaude-code-linux-x64	https://www.npmjs.com/search?q=anthropic	AI Utility / Integration	Anthropic	Claude	2	True	False	False	False	False	high	2026-06-01T02:14:11Z	3
npm-package-@anthropic-ai/claude-code	@anthropic-ai/claude-code	@anthropic-ai	claude-code	Use Claude, Anthropic's AI assistant, right from your terminal. Claude can understand your codebase,... [72 more chars]	2.1.159	SEE LICENSE IN README.md	Anthropic	wolffiex	14	2026-05-31T17:50:06.818Z	2025-02-24T18:03:40.929Z	461	0	True	False		8661516	33327160	306.419	anthropic	claude	2	anthropic	3	45		0	0	0	0	1	True	False	False	>=18.0.0			https://github.com/anthropics/claude-code	github.com	https://github.com/anthropics/claude-code/issues	https://www.npmjs.com/package/@anthropic-ai/claude-code	https://registry.npmjs.org/%40anthropic-ai%2Fclaude-code	https://www.npmjs.com/search?q=anthropic	Developer Tooling / CLI	Anthropic	Claude	2	True	False	False	False	False	high	2026-06-01T02:14:11Z	4
npm-package-@anthropic-ai/claude-agent-sdk	@anthropic-ai/claude-agent-sdk	@anthropic-ai	claude-agent-sdk	SDK for building AI agents with Claude Code's capabilities. Programmatically interact with Claude to... [90 more chars]	0.3.159	SEE LICENSE IN README.md	Anthropic	wolffiex	14	2026-05-31T17:50:04.652Z	2025-09-27T16:07:20.263Z	246	0	True	False		7066332	25180998	316.142	anthropic	claude	2	anthropic	1	31	ai	agent	sdk	claude	anthropic	automation	code-generation	7	0	0	3	0	False	True	True	>=18.0.0	https://github.com/anthropics/claude-agent-sdk-typescript.git	github.com	https://github.com/anthropics/claude-agent-sdk-typescript	github.com	https://github.com/anthropics/claude-agent-sdk-typescript/issues	https://www.npmjs.com/package/@anthropic-ai/claude-agent-sdk	https://registry.npmjs.org/%40anthropic-ai%2Fclaude-agent-sdk	https://www.npmjs.com/search?q=anthropic	Agent Framework	Anthropic	Claude	2	True	False

---

Dataset Structure

Total rows: 60

Total columns: 56

Splits

• data: 60 rows

Data Files

• data: data/data.parquet

Data Dictionary

The table below describes the columns included in this dataset.

column	pandas_dtype	dataset_type	description
record_id	object	string	Text column. Stable unique identifier for the dataset row derived from the npm package name.
package_name	object	string	Text column. Canonical npm package name.
package_scope	object	string	Text column. Scope prefix for scoped packages, such as @modelcontextprotocol, when present.
unscoped_name	object	string	Text column. Package name segment after the scope separator or the full name for unscoped packages.
package_description	object	string	Text column. Public package description from npm metadata.
latest_version	object	string	Text column. Latest version string reported by the npm registry.
license_name	object	string	Text column. License identifier exposed by the package metadata when available.
author_name	object	string	Text column. Author name parsed from npm package metadata when available.
publisher_name	object	string	Text column. Publisher username associated with the latest published version when available.
maintainer_count	int64	integer	Whole-number numeric column. Count of maintainers exposed in the registry payload.
published_at	object	datetime-like string	Stored as text, but values appear to represent dates or timestamps. Timestamp of the latest published version.
first_published_at	object	datetime-like string	Stored as text, but values appear to represent dates or timestamps. Timestamp of the package creation time in the npm registry.
package_age_days	int64	integer	Whole-number numeric column. Days between first publication and collection time.
days_since_publish	int64	integer	Whole-number numeric column. Days between the latest published version and collection time.
is_recently_published_90d	bool	boolean	True/false column. Boolean flag indicating the latest package version was published in the last 90 days.
is_deprecated	bool	boolean	True/false column. Boolean flag indicating the latest version is marked deprecated.
deprecated_message	object	empty	Column contains no non-null values. Deprecation message for the latest version when present.
weekly_downloads	int64	integer	Whole-number numeric column. Official npm last-week download count.
monthly_downloads	int64	integer	Whole-number numeric column. Official npm last-month download count.
search_score	float64	float	Decimal numeric column. Maximum npm search score observed across the matched search terms.
matched_search_terms	object	string	Text column. Pipe-delimited list of search terms that returned this package.
matched_search_term_count	int64	integer	Whole-number numeric column. Number of distinct search terms that returned this package.
primary_search_term	object	string	Text column. First matched search term used as the primary provenance label for the package.
best_search_rank	int64	integer	Whole-number numeric column. Best observed search rank across all matched search terms.
mean_search_rank	float64	float	Decimal numeric column. Average search rank across all matched search terms.
keywords	object	string	Text column. Pipe-delimited package keywords from npm metadata.
keyword_count	int64	integer	Whole-number numeric column. Number of package keywords.
dependency_count	int64	integer	Whole-number numeric column. Count of production dependencies in the latest package version.
dev_dependency_count	int64	integer	Whole-number numeric column. Count of development dependencies in the latest package version.
peer_dependency_count	int64	integer	Whole-number numeric column. Count of peer dependencies in the latest package version.
bin_command_count	int64	integer	Whole-number numeric column. Count of CLI binary entries exposed by the package.
has_cli_bin	bool	boolean	True/false column. Boolean flag indicating the package exposes at least one CLI command.
has_typescript_types	bool	boolean	True/false column. Boolean flag indicating the package exposes built-in TypeScript typings.
has_exports	bool	boolean	True/false column. Boolean flag indicating the package defines an exports field.
engines_node	object	string	Text column. Declared Node.js engine range from the latest version when present.
repository_url	object	string	Text column. Repository URL from npm metadata when present.
repository_domain	object	string	Text column. Repository host domain parsed from repository_url.
homepage_url	object	string	Text column. Homepage URL from npm metadata when present.
homepage_domain	object	string	Text column. Homepage host domain parsed from homepage_url.
bugs_url	object	string	Text column. Issue tracker or bugs URL from npm metadata when present.
npm_package_url	object	string	Text column. Public npm package page URL.
registry_url	object	string	Text column. Public npm registry JSON endpoint used for enrichment.
source_search_url	object	string	Text column. Public npm search URL for the primary search term.
ecosystem_segment	object	string	Text column. Heuristic market segment label derived from package metadata and keyword signals.
provider_mentions	object	string	Text column. Pipe-delimited provider and framework mentions detected in package metadata.
provider_count	int64	integer	Whole-number numeric column. Count of detected provider and framework mentions.
has_core_ai_signal	bool	boolean	True/false column. Boolean flag indicating core AI or LLM terminology appears in the package metadata.
has_mcp_signal	bool	boolean	True/false column. Boolean flag indicating MCP terminology appears in the package metadata.
has_agent_signal	bool	boolean	True/false column. Boolean flag indicating agent or orchestration terminology appears in the package metadata.
has_sdk_signal	bool	boolean	True/false column. Boolean flag indicating SDK, client, adapter, or integration terminology appears in the package metadata.
has_commercial_terms	bool	boolean	True/false column. Boolean flag indicating commercialization-oriented terms appear in the package metadata.
commercial_readiness	object	string	Text column. Heuristic commercialization label derived from popularity, freshness, and deprecation status.
last_collected_at	object	datetime-like string	Stored as text, but values appear to represent dates or timestamps. UTC timestamp when the row was collected.
monthly_download_rank	Int64	integer	Whole-number numeric column. Descending rank of monthly downloads within this dataset build.
weekly_download_rank	Int64	integer	Whole-number numeric column. Descending rank of weekly downloads within this dataset build.
downloads_tier	object	string	Text column. Bucketed popularity tier derived from monthly downloads.

Intended Use

This dataset is intended for research, experimentation, analysis, and model prototyping.

Loading the Dataset

import os
from datasets import load_dataset

HUGGINGFACE_API_KEY_KARMANE = os.environ.get("HUGGINGFACE_API_KEY_KARMANE")

dataset = load_dataset(
    "Karmane/npm-ai-agent-mcp-packages-enriched",
    token=HUGGINGFACE_API_KEY_KARMANE,
)

print(dataset)
print(dataset[list(dataset.keys())[0]][0])

# getting the DataFrame itself
# df = dataset[list(dataset.keys())[0]].to_pandas()

---

Karmane. (2025). npm AI Agent and MCP Packages Enriched Dataset. Hugging Face. https://huggingface.co/datasets/Karmane/npm-ai-agent-mcp-packages-enriched