fix tool issues and add quickstart

README.md

@@ -8,41 +8,139 @@ app_file: app.py
 pinned: false
 ---
 
-# Weights & Biases MCP Server on HuggingFace Spaces
+<p align="center">
+  <picture>
+    <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/wandb/wandb/main/assets/logo-dark.svg">
+    <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/wandb/wandb/main/assets/logo-light.svg">
+    <img src="https://raw.githubusercontent.com/wandb/wandb/main/assets/logo-light.svg" width="600" alt="Weights & Biases">
+  </picture>
+</p>
 
-A [Model Context Protocol (MCP)](https://modelcontextprotocol.io/) server for querying [Weights & Biases](https://www.wandb.ai/) data, hosted on HuggingFace Spaces.
+# Weights & Biases MCP Server
 
-This server allows MCP clients to:
-- 📊 Query W&B Models runs and sweeps
-- 🔍 Query W&B Weave traces, evaluations and datasets  
-- 🤖 Query [wandbot](https://github.com/wandb/wandbot), the W&B support agent
-- 📝 Write text and charts to W&B Reports
+A Model Context Protocol (MCP) server that provides seamless access to [Weights & Biases](https://www.wandb.ai/) for ML experiments and agent applications.
 
-## 🚀 Quick Start - Use This Server Directly!
+## Example Use Cases
 
-### No Setup Required! 🎉
+### 1. 🔍 Analyze ML Experiments
+```
+"Show me the top 5 runs with the highest accuracy from my wandb-smle/hiring-agent-demo-public project and create a report comparing their hyperparameters"
+```
+The MCP server queries W&B runs, compares metrics, and generates a shareable report with visualizations.
 
-You can use this server immediately with your own W&B API key:
+### 2. 🐛 Debug LLM Applications  
+```
+"Find all failed OpenAI chat traces in my weave project from the last 24 hours and analyze their error patterns"
+```
+The server retrieves Weave traces, filters by status, and provides detailed error analysis for debugging.
 
-1. **Get your W&B API key** from [wandb.ai/authorize](https://wandb.ai/authorize)
-2. **Configure your MCP client** with:
-   - **Server URL**: `https://niware-wandb-mcp-server.hf.space/mcp`
-   - **Authentication**: Your W&B API key as Bearer token
-3. **Start querying** your W&B data!
+### 3. 📊 Evaluate Model Performance
+```
+"Compare the F1 scores across all evaluations in my RAG pipeline and identify which prompts performed best"
+```
+The server queries Weave evaluations, aggregates scores, and highlights top-performing configurations.
+
+### 4. 🤖 Get Expert Help with W&B/Weave
+```
+"How do I implement custom metrics in Weave evaluations? Show me an example with async scorers"
+```
+The integrated [wandbot](https://github.com/wandb/wandbot) support agent provides detailed answers, code examples, and debugging assistance for any W&B or Weave-related questions.
 
-That's it! No server configuration, no duplication needed. Each user provides their own API key, ensuring secure access to their own W&B projects.
+## Deployment Options
 
-## 🖥️ Using with MCP Clients
+This MCP server can be deployed in three ways:
 
-### Mistral LeChat
-1. Go to Settings → Custom MCP Connectors
-2. Add a new connector:
-   - **Server URL**: `https://niware-wandb-mcp-server.hf.space/mcp`
-   - **Authentication**: Choose "API Key Authentication"
-   - **API Key**: Your W&B API key from [wandb.ai/authorize](https://wandb.ai/authorize)
+### 🌐 Option 1: Use the Hosted Server (Recommended)
+
+Use our publicly hosted server on Hugging Face Spaces - no installation needed!
+
+**Server URL:** `https://niware-wandb-mcp-server.hf.space/mcp`
+
+Configure your MCP client to connect to the hosted server with your W&B API key as authentication. See the [Client Configuration](#mcp-client-configuration-for-hosted-server) section below for details.
+
+### 💻 Option 2: Local Development (STDIO)
+
+Run the server locally with direct stdio communication - best for development and testing.
+
+### 🔌 Option 3: Self-Hosted HTTP Server
+
+Deploy your own HTTP server with API key authentication - great for team deployments or custom infrastructure.
+
+---
+
+## Installation
+
+### For Hosted Server Users
+
+No installation needed! Skip to [Client Configuration](#mcp-client-configuration-for-hosted-server).
+
+### For Local Installation
+
+These instructions are for running the MCP server locally (Options 2 & 3).
+
+### Prerequisites
+
+#### 1. Install UV Package Manager
+
+UV is required to run the MCP server. Install it using one of these methods:
+
+**macOS/Linux:**
+```bash
+curl -LsSf https://astral.sh/uv/install.sh | sh
+```
+
+**macOS (Homebrew):**
+```bash
+brew install uv
+```
+
+**Windows:**
+```powershell
+powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
+```
+
+#### 2. Get Your W&B API Key
+
+You'll need a Weights & Biases API key. Get yours at: [https://wandb.ai/authorize](https://wandb.ai/authorize)
+
+Configure your API key using one of these methods (first one recommended to have the other default parameters too):
+
+1. **`.env` file** in your project (copy from `env.example`):
+   ```bash
+   cp env.example .env
+   # Edit .env and add your API key
+   ```
+
+2. **`.netrc` file**:
+   ```bash
+   uvx wandb login
+   ```
+
+3. **Environment variable** (recommended):
+   ```bash
+   export WANDB_API_KEY=your-api-key
+   ```
+
+4. **Command-line argument**:
+   ```bash
+   wandb_mcp_server --wandb-api-key your-api-key
+   ```
+
+#### 3. Environment Configuration (Optional)
+
+The server includes [wandbot](https://github.com/wandb/wandbot) support for answering W&B/Weave questions. **wandbot works out-of-the-box without any configuration!** It uses the default public endpoint automatically.
+
+See `env.example` for optional configuration like custom wandbot instances or other advanced settings.
+
+### MCP Client Configuration for Hosted Server
+
+To use the hosted server, configure your MCP client with the following settings:
+
+<details>
+<summary><b>🖱️ Cursor IDE (Hosted Server)</b></summary>
+
+Add to `.cursor/mcp.json` or `~/.cursor/mcp.json`:
 
-### Cursor
-Add to `.cursor/mcp.json` (project) or `~/.cursor/mcp.json` (global):
 ```json
 {
   "mcpServers": {
@@ -58,81 +156,419 @@ Add to `.cursor/mcp.json` (project) or `~/.cursor/mcp.json` (global):
 }
 ```
 
-### Claude Desktop / Claude Code
-Configure in your MCP settings with the server URL and Bearer token authentication.
+Replace `YOUR_WANDB_API_KEY` with your actual W&B API key from [wandb.ai/authorize](https://wandb.ai/authorize).
+</details>
+
+<details>
+<summary><b>🎨 Mistral LeChat (Hosted Server)</b></summary>
+
+1. Go to LeChat Settings → Custom MCP Connectors
+2. Click "Add MCP Connector"
+3. Configure with:
+   - **Server URL**: `https://niware-wandb-mcp-server.hf.space/mcp`
+   - **Authentication**: Choose "API Key Authentication"
+   - **Token**: Enter your W&B API key
+</details>
+
+### MCP Client Setup for Local Server
+
+Choose your MCP client from the options below for local server setup:
+
+<details>
+<summary><b>🖱️ Cursor IDE</b></summary>
+
+**Quick Install (Project-specific):**
+```bash
+uvx --from git+https://github.com/wandb/wandb-mcp-server -- add_to_client --config_path .cursor/mcp.json && uvx wandb login
+```
+
+**Quick Install (Global):**
+```bash
+uvx --from git+https://github.com/wandb/wandb-mcp-server -- add_to_client --config_path ~/.cursor/mcp.json && uvx wandb login
+```
+
+<details>
+<summary>Manual Configuration</summary>
+
+Add to `.cursor/mcp.json` or `~/.cursor/mcp.json`:
+
+```json
+{
+  "mcpServers": {
+    "wandb": {
+      "command": "uvx",
+      "args": [
+        "--from",
+        "git+https://github.com/wandb/wandb-mcp-server",
+        "wandb_mcp_server"
+      ],
+      "env": {
+        "WANDB_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
+```
+</details>
+</details>
+
+<details>
+<summary><b>🌊 Windsurf IDE</b></summary>
+
+**Quick Install:**
+```bash
+uvx --from git+https://github.com/wandb/wandb-mcp-server -- add_to_client --config_path ~/.codeium/windsurf/mcp_config.json && uvx wandb login
+```
+
+<details>
+<summary>Manual Configuration</summary>
+
+Add to `~/.codeium/windsurf/mcp_config.json`:
+
+```json
+{
+  "mcpServers": {
+    "wandb": {
+      "command": "uvx",
+      "args": [
+        "--from",
+        "git+https://github.com/wandb/wandb-mcp-server",
+        "wandb_mcp_server"
+      ],
+      "env": {
+        "WANDB_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
+```
+</details>
+</details>
+
+<details>
+<summary><b>💬 Gemini</b></summary>
+**Quick Install:**
+Uses the `.gemini-extension.json` in this repo's root:
 
-## 🔧 Available MCP Tools
+```bash
+gemini extensions install https://github.com/wandb/wandb-mcp-server
+```
+
+**Then set your API key (choose one):**
+```bash
+# Option 1: Export API key directly
+export WANDB_API_KEY=your-api-key
 
-### W&B Models
-- **`query_wandb_tool`**: Execute GraphQL queries against W&B experiment tracking data
+# Option 2: Use wandb login (opens browser)
+uvx wandb login
+```
+<details>
+<summary>Manual Configuration</summary>
+Create `gemini-extension.json` in your project root (use `--path=path/to/gemini-extension.json` to add local folder):
 
-### W&B Weave  
-- **`query_weave_traces_tool`**: Query Weave evaluations and traces with filtering and pagination
-- **`count_weave_traces_tool`**: Count traces matching filters without returning data
+```json
+{
+    "name": "Weights and Biases MCP Server",
+    "version": "0.1.0",
+    "mcpServers": {
+        "wandb": {
+            "command": "uv",
+            "args": [
+                "run", 
+                "--directory",
+                "/path/to/wandb-mcp-server",
+                "wandb_mcp_server",
+                "--transport",
+                "stdio"
+            ],
+            "env": {
+                "WANDB_API_KEY": "$WANDB_API_KEY"
+            }
+        }
+    }
+}
+```
+</details>
 
-### Support & Reports
-- **`query_wandb_support_bot`**: Get help from wandbot, the W&B support agent
-- **`create_wandb_report_tool`**: Create W&B Reports with markdown and visualizations
-- **`query_wandb_entity_projects`**: List available W&B entities and projects
+Note: Replace `/path/to/wandb-mcp-server` with your installation path.
+</details>
 
-## 📝 Example Queries
+<details>
+<summary><b>🤖 Claude Desktop</b></summary>
 
+**Quick Install:**
+```bash
+uvx --from git+https://github.com/wandb/wandb-mcp-server -- add_to_client --config_path "~/Library/Application Support/Claude/claude_desktop_config.json" && uvx wandb login
 ```
-How many openai.chat traces are in my wandb-team/my-project weave project?
+
+<details>
+<summary>Manual Configuration</summary>
+
+Add to `~/Library/Application Support/Claude/claude_desktop_config.json` (macOS) or `%APPDATA%\Claude\claude_desktop_config.json` (Windows):
+
+```json
+{
+  "mcpServers": {
+    "wandb": {
+      "command": "uvx",
+      "args": [
+        "--from",
+        "git+https://github.com/wandb/wandb-mcp-server",
+        "wandb_mcp_server"
+      ],
+      "env": {
+        "WANDB_API_KEY": "your-api-key"
+      }
+    }
+  }
+}
 ```
+</details>
+</details>
+
+<details>
+<summary><b>💻 Claude Code</b></summary>
 
+**Quick Install:**
+```bash
+claude mcp add wandb -- uvx --from git+https://github.com/wandb/wandb-mcp-server wandb_mcp_server && uvx wandb login
 ```
-Show me the latest 10 runs from my experiment tracking project and create a report with the results.
+
+**With API Key:**
+```bash
+claude mcp add wandb -e WANDB_API_KEY=your-api-key -- uvx --from git+https://github.com/wandb/wandb-mcp-server wandb_mcp_server
 ```
+</details>
+
+<details>
+<summary><b>🌐 ChatGPT, LeChat, Claude</b></summary>
+Try our hosted public version: [HF Spaces](https://huggingface.co/spaces/NiWaRe/wandb-mcp-server)
+
+This version allows you to configure your WANDB_API_KEY directly in the interface to access your own projects or to work with all publich projects otherwise. Follow the instructions in the space to add it to LeChat, ChatGPT, or Claude. We'll have an official hosted version soon.
+</details>
+
+## Available Tools
+
+The server provides the following MCP tools:
+
+### W&B Models Tools
+- **`query_wandb_tool`** - Execute GraphQL queries against W&B experiment tracking data (runs, sweeps, artifacts)
+
+### Weave Tools  
+- **`query_weave_traces_tool`** - Query LLM traces and evaluations with filtering and pagination
+- **`count_weave_traces_tool`** - Efficiently count traces without returning data
+
+### Support & Reporting
+- **`query_wandb_support_bot`** - Get help from [wandbot](https://github.com/wandb/wandbot), our RAG-powered technical support agent that can answer any W&B/Weave questions, help debug issues, and provide code examples (works out-of-the-box, no configuration needed!)
+- **`create_wandb_report_tool`** - Create W&B Reports with markdown and visualizations
+- **`query_wandb_entity_projects`** - List available entities and projects
+
+## Usage Tips
+
+### Be Specific About Projects
+Always specify the W&B entity and project name in your queries:
+
+✅ **Good:** "Show traces from wandb-team/my-project"  
+❌ **Bad:** "Show my traces"
+
+### Avoid Overly Broad Questions
+Be specific to get better results:
+
+✅ **Good:** "What eval had the highest F1 score in the last week?"  
+❌ **Bad:** "What's my best evaluation?"
+
+### Verify Complete Data Retrieval
+When analyzing performance across multiple runs, ask the LLM to confirm it retrieved all available data to ensure comprehensive analysis.
+
+## Self-Hosting Guide
+
+### Deploy to Hugging Face Spaces
+
+Deploy your own instance of the W&B MCP Server on Hugging Face Spaces:
+
+1. **Fork this repository** or clone it locally
+2. **Create a new Space on Hugging Face:**
+   - Go to [huggingface.co/spaces](https://huggingface.co/spaces)
+   - Click "Create new Space"
+   - Choose "Docker" as the SDK
+   - Set visibility as needed
+
+3. **Push the code to your Space:**
+   ```bash
+   git remote add hf-space https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME
+   git push hf-space main
+   ```
+
+4. **Your server will be available at:**
+   ```
+   https://YOUR_USERNAME-YOUR_SPACE_NAME.hf.space/mcp
+   ```
+
+See [HUGGINGFACE_DEPLOYMENT.md](HUGGINGFACE_DEPLOYMENT.md) for detailed deployment instructions.
+
+### Run Local HTTP Server
 
+Run the server locally with HTTP transport for development or testing:
+
+```bash
+# Install dependencies
+pip install -r requirements.txt
+
+# Run with authentication (recommended)
+python app.py
+
+# Or run without authentication (development only)
+MCP_AUTH_DISABLED=true python app.py
+```
+
+The server will be available at `http://localhost:7860/mcp`
+
+**Authentication:** See [AUTH_README.md](AUTH_README.md) for details on Bearer token authentication.
+
+### File Structure for Deployment
+
+```
+wandb-mcp-server/
+├── app.py                    # HF Spaces/HTTP server entry point
+├── Dockerfile                # Container configuration for HF Spaces
+├── requirements.txt          # Python dependencies for HTTP deployment
+├── index.html               # Landing page for web interface
+├── AUTH_README.md           # Authentication documentation
+├── HUGGINGFACE_DEPLOYMENT.md # HF Spaces deployment guide
+├── src/
+│   └── wandb_mcp_server/
+│       ├── server.py        # Core MCP server (STDIO & HTTP)
+│       ├── auth.py          # Bearer token authentication
+│       └── mcp_tools/       # Tool implementations
+└── pyproject.toml           # Package configuration for local/pip install
+```
+
+## Advanced Configuration
+
+### Enabling Weave Tracing for MCP Operations
+
+Track all MCP tool calls using [Weave's MCP integration](https://weave-docs.wandb.ai/guides/integrations/mcp):
+
+```bash
+# Enable Weave tracing for MCP operations
+export WEAVE_DISABLED=false
+export MCP_LOGS_WANDB_ENTITY=your-entity
+export MCP_LOGS_WANDB_PROJECT=mcp-logs
+
+# Optional: trace list operations
+export MCP_TRACE_LIST_OPERATIONS=true
+```
+
+This provides detailed observability into tool calls, resource access, and prompt generation across your MCP system.
+
+### Logging Configuration
+
+Control server logging with environment variables:
+
+```bash
+# Server log level
+export MCP_SERVER_LOG_LEVEL=INFO  # DEBUG, INFO, WARNING, ERROR
+
+# W&B/Weave output control
+export WANDB_SILENT=False  # Show W&B output
+export WEAVE_SILENT=False  # Show Weave output
+
+# Debug mode
+export WANDB_DEBUG=true    # Verbose W&B logging
 ```
-What's the best performing model in my latest sweep? Plot the results.
+
+### Transport Options
+
+#### STDIO Transport (Default for Local Development)
+For local development where the MCP client and server run on the same machine:
+```bash
+wandb_mcp_server --transport stdio
+# Or with UV:
+uvx --from git+https://github.com/wandb/wandb-mcp-server wandb_mcp_server
 ```
+- Requires W&B API key in environment
+- Direct communication via stdin/stdout
+- Best for local IDE integrations (Cursor, Windsurf, etc.)
+
+#### HTTP Transport (For Remote Access)
+For remote access, web applications, or hosted deployments:
+```bash
+# Using the FastAPI app (recommended)
+python app.py  # Runs on port 7860 by default
 
-## 💻 Run Locally (Optional)
+# Or using the CLI
+wandb_mcp_server --transport http --host 0.0.0.0 --port 8080
+```
+- Clients provide W&B API key as Bearer token
+- Supports authentication middleware
+- Uses Server-Sent Events (SSE) for streaming
+- Ideal for hosted deployments and web clients
 
-Want to run your own instance or develop locally? Check out the main repository:
+### Running from Source
 
 ```bash
-# Install and run from source
 git clone https://github.com/wandb/wandb-mcp-server
 cd wandb-mcp-server
-uv run src/wandb_mcp_server/server.py --transport http
+wandb login
+uv run src/wandb_mcp_server/server.py
 ```
 
-See the [GitHub repository](https://github.com/wandb/wandb-mcp-server) for installation instructions for various MCP clients.
+## Troubleshooting
 
-## 🔍 Troubleshooting
+### Error: spawn uv ENOENT
 
-### Connection Issues?
-1. **Verify your API key**: Ensure it's correctly copied from [wandb.ai/authorize](https://wandb.ai/authorize)
-2. **Check the endpoint**: Must include `/mcp` suffix: `https://niware-wandb-mcp-server.hf.space/mcp`
-3. **Headers required**: Include both `Authorization` and `Accept` headers as shown above
+If `uv` cannot be found:
 
-### Query Tips
-- Always specify your W&B entity and project name in queries
-- Be specific rather than overly broad in your questions
-- Verify you have access to the projects you're querying
+1. Reinstall UV:
+   ```bash
+   curl -LsSf https://astral.sh/uv/install.sh | sh
+   ```
 
-## 🔐 Security
+2. Create a system-wide symlink:
+   ```bash
+   sudo ln -s ~/.local/bin/uv /usr/local/bin/uv
+   ```
 
-- **Your API key is never stored** - It's only used transiently for your requests
-- **Each user is isolated** - Your key only accesses your W&B data
-- **No server configuration needed** - The server doesn't have its own W&B access
-- **Industry-standard Bearer tokens** - Same pattern as GitHub, OpenAI, etc.
+3. Restart your application/IDE
 
-## 📚 Resources
+### Authentication Issues
 
-- [Model Context Protocol Documentation](https://modelcontextprotocol.io/)
-- [Weights & Biases Documentation](https://docs.wandb.ai/)
-- [W&B Weave Documentation](https://weave-docs.wandb.ai/)
-- [Source Code Repository](https://github.com/wandb/wandb-mcp-server)
-- [Get Your W&B API Key](https://wandb.ai/authorize)
+Verify W&B authentication:
+```bash
+uvx wandb login
+```
 
-## 📄 License
+Or check if your API key is set:
+```bash
+echo $WANDB_API_KEY
+```
 
-This project is licensed under the MIT License. See the [GitHub repository](https://github.com/wandb/wandb-mcp-server) for details.
+## Testing
 
----
+Run integration tests with LLM providers:
+
+```bash
+# Set API key in .env
+echo "ANTHROPIC_API_KEY=your-key" >> .env
+
+# Run specific test file
+uv run pytest -s -n 10 tests/test_query_wandb_gql.py
+
+# Debug single test
+pytest -s -n 1 "tests/test_query_weave_traces.py::test_query_weave_trace[sample_name]" -v --log-cli-level=DEBUG
+```
+
+## Contributing
+
+We welcome contributions! Please see our [Contributing Guide](CONTRIBUTING.md) for details.
+
+## License
+
+This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
+
+## Support
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+- [W&B Documentation](https://docs.wandb.ai)
+- [Weave Documentation](https://weave-docs.wandb.ai)  
+- [GitHub Issues](https://github.com/wandb/wandb-mcp-server/issues)
+- [W&B Community Forum](https://community.wandb.ai)

gemini-extension.json

@@ -0,0 +1,20 @@
+{
+    "name": "wandb-weave",
+    "version": "0.1.0",
+    "mcpServers": {
+        "wandb-weave": {
+            "command": "uv",
+            "args": [
+                "run", 
+                "--directory",
+                "/Users/niware_wb/Documents/code_projects/mcp_experiments/wandb-mcp-server",
+                "wandb_mcp_server",
+                "--transport",
+                "stdio"
+            ],
+            "env": {
+                "WANDB_API_KEY": "<your-api-key>"
+            }
+        }
+    }
+}

index.html

@@ -230,13 +230,69 @@
         </div>
 
         <div class="setup-steps" style="background: #f0fdf4; border-color: #86efac;">
-            <h3 style="color: #166534;">✅ Ready to Use!</h3>
-            <ol>
-                <li>Get your W&B API key from <a href="https://wandb.ai/authorize" target="_blank">wandb.ai/authorize</a></li>
-                <li>Configure your MCP client with the endpoint above</li>
-                <li>Provide your API key as Bearer token or API Key authentication</li>
-            </ol>
-            <p style="margin-top: 15px; color: #166534;"><strong>No server configuration needed!</strong> Each user provides their own API key for secure access to their W&B data.</p>
+            <h3 style="color: #166534;">🚀 Quick Setup</h3>
+            
+            <div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); gap: 20px; margin: 20px 0;">
+                
+                <div style="background: white; padding: 20px; border-radius: 8px; border-left: 4px solid #4299e1;">
+                    <h4 style="margin: 0 0 10px 0; color: #2d3748;">📝 Cursor</h4>
+                    <ul style="margin: 0; padding-left: 20px; font-size: 0.9em;">
+                        <li>Add to <code>~/.cursor/mcp.json</code>:</li>
+                    </ul>
+                    <pre style="background: #1a202c; color: #e2e8f0; padding: 10px; border-radius: 4px; font-size: 0.8em; margin: 10px 0; overflow-x: auto;">"wandb": {
+  "transport": "http", 
+  "url": "https://niware-wandb-mcp-server.hf.space/mcp",
+  "headers": {
+    "Authorization": "Bearer YOUR_WANDB_API_KEY",
+    "Accept": "application/json, text/event-stream"
+  }
+}</pre>
+                </div>
+
+                <div style="background: white; padding: 20px; border-radius: 8px; border-left: 4px solid #38b2ac;">
+                    <h4 style="margin: 0 0 10px 0; color: #2d3748;">🖥️ Claude Desktop</h4>
+                    <ul style="margin: 0; padding-left: 20px; font-size: 0.9em;">
+                        <li>Add to <code>claude_desktop_config.json</code>:</li>
+                    </ul>
+                    <pre style="background: #1a202c; color: #e2e8f0; padding: 10px; border-radius: 4px; font-size: 0.8em; margin: 10px 0; overflow-x: auto;">{
+  "mcpServers": {
+    "wandb-stdio": {
+      "command": "uvx",
+      "args": ["--from", "git+https://github.com/wandb/wandb-mcp-server.git", "wandb-mcp-server"],
+      "env": {
+        "WANDB_API_KEY": "YOUR_WANDB_API_KEY"
+      }
+    }
+  }
+}</pre>
+                </div>
+
+                <div style="background: white; padding: 20px; border-radius: 8px; border-left: 4px solid #f56565;">
+                    <h4 style="margin: 0 0 10px 0; color: #2d3748;">💬 LeChat</h4>
+                    <ul style="margin: 0; padding-left: 20px; font-size: 0.9em;">
+                        <li>MCP Server URL: <code>https://niware-wandb-mcp-server.hf.space/mcp</code></li>
+                        <li>Auth: Bearer token</li>
+                        <li>Token: Your W&B API key</li>
+                    </ul>
+                </div>
+
+                <div style="background: white; padding: 20px; border-radius: 8px; border-left: 4px solid #9f7aea;">
+                    <h4 style="margin: 0 0 10px 0; color: #2d3748;">🤖 Gemini CLI</h4>
+                    <ul style="margin: 0; padding-left: 20px; font-size: 0.9em;">
+                        <li>Install extension:</li>
+                    </ul>
+                    <pre style="background: #1a202c; color: #e2e8f0; padding: 10px; border-radius: 4px; font-size: 0.8em; margin: 10px 0; overflow-x: auto;">gemini extensions install https://github.com/wandb/wandb-mcp-server</pre>
+                    <ul style="margin: 10px 0 0 0; padding-left: 20px; font-size: 0.9em;">
+                        <li>Then set API key (choose one):</li>
+                        <li style="margin-left: 15px;">• <code>export WANDB_API_KEY=your-key</code></li>
+                        <li style="margin-left: 15px;">• <code>uvx wandb login</code> (opens browser)</li>
+                    </ul>
+                </div>
+            </div>
+            
+            <p style="margin-top: 15px; color: #166534; text-align: center;">
+                <strong>Get your API key:</strong> <a href="https://wandb.ai/authorize" target="_blank" style="color: #166534;">wandb.ai/authorize</a>
+            </p>
         </div>
 
         <div class="examples">

pyproject.toml

@@ -14,8 +14,6 @@ dependencies = [
     "wandb-workspaces>=0.1.12",
     "networkx>=3.4.2",
     "requests>=2.31.0",
-    "fastapi>=0.104.0",
-    "uvicorn>=0.24.0",
 ]
 
 [build-system]
@@ -33,6 +31,10 @@ test = [
     "pytest-xdist>=3.6.1",
     "pytest-asyncio>=0.26.0",
 ]
+http = [
+    "fastapi>=0.104.0",
+    "uvicorn>=0.24.0",
+]
 
 [tool.hatch.build.targets.wheel]
 packages = ["src/wandb_mcp_server"]

src/wandb_mcp_server/mcp_tools/list_wandb_entities_projects.py

@@ -94,9 +94,9 @@ def list_entity_projects(entity: str | None = None) -> dict[str, list[dict[str,
                 "entity": project.entity,
                 "description": getattr(project, "description", None),
                 "visibility": getattr(project, "visibility", None),
-                "created_at": project.created_at,
-                "updated_at": project.updated_at,
-                "tags": project.tags,
+                "created_at": getattr(project, "created_at", None),
+                "updated_at": getattr(project, "updated_at", None),
+                "tags": getattr(project, "tags", []),
             }
             projects_data.append(project_dict)
 

src/wandb_mcp_server/mcp_tools/query_wandbot.py

@@ -34,7 +34,7 @@ str
 
 
 def query_wandbot_api(question: str) -> Dict[str, Any]:
-    wandbot_base_url = os.getenv("WANDBOT_BASE_URL", "https://morg--wandbot-api-wandbotapi-serve.modal.run")
+    wandbot_base_url = os.getenv("WANDBOT_BASE_URL", "https://weightsandbiases-wandbot--wandbot-api-wandbotapi-serve.modal.run")
     QUERY_ENDPOINT = f"{wandbot_base_url}/chat/query"
     STATUS_ENDPOINT = f"{wandbot_base_url}/status"
     QUERY_TIMEOUT_SECONDS = 40

src/wandb_mcp_server/server.py

@@ -276,14 +276,14 @@ def register_tools(mcp_instance: FastMCP) -> None:
     async def query_weave_traces_tool(
         entity_name: str,
         project_name: str,
-        filters: Dict = {},
+        filters: Dict[str, Any] = {},
         sort_by: str = "started_at",
         sort_direction: str = "desc",
         limit: int = 10000000,
         include_costs: bool = True,
         include_feedback: bool = True,
-        columns: list = [],
-        expand_columns: list = [],
+        columns: List[str] = [],
+        expand_columns: List[str] = [],
         truncate_length: int = 200,
         return_full_data: bool = False,
         metadata_only: bool = False,

uv.lock

@@ -2092,7 +2092,6 @@ name = "wandb-mcp-server"
 version = "0.1.0"
 source = { editable = "." }
 dependencies = [
-    { name = "fastapi" },
     { name = "httpx" },
     { name = "mcp", extra = ["cli"] },
     { name = "networkx" },
@@ -2101,13 +2100,16 @@ dependencies = [
     { name = "requests" },
     { name = "simple-parsing" },
     { name = "tiktoken" },
-    { name = "uvicorn" },
     { name = "wandb" },
     { name = "wandb-workspaces" },
     { name = "weave" },
 ]
 
 [package.optional-dependencies]
+http = [
+    { name = "fastapi" },
+    { name = "uvicorn" },
+]
 test = [
     { name = "anthropic" },
     { name = "litellm" },
@@ -2123,7 +2125,7 @@ dev = [
 [package.metadata]
 requires-dist = [
     { name = "anthropic", marker = "extra == 'test'", specifier = ">=0.50.0" },
-    { name = "fastapi", specifier = ">=0.104.0" },
+    { name = "fastapi", marker = "extra == 'http'", specifier = ">=0.104.0" },
     { name = "httpx", specifier = ">=0.28.1" },
     { name = "litellm", marker = "extra == 'test'", specifier = ">=1.67.2" },
     { name = "mcp", extras = ["cli"], specifier = ">=1.0.0" },
@@ -2135,12 +2137,12 @@ requires-dist = [
     { name = "requests", specifier = ">=2.31.0" },
     { name = "simple-parsing", specifier = ">=0.1.7" },
     { name = "tiktoken", specifier = ">=0.9.0" },
-    { name = "uvicorn", specifier = ">=0.24.0" },
+    { name = "uvicorn", marker = "extra == 'http'", specifier = ">=0.24.0" },
     { name = "wandb", specifier = ">=0.19.8" },
     { name = "wandb-workspaces", specifier = ">=0.1.12" },
     { name = "weave", specifier = ">=0.51.56" },
 ]
-provides-extras = ["test"]
+provides-extras = ["test", "http"]
 
 [package.metadata.requires-dev]
 dev = [{ name = "ruff", specifier = ">=0.11.12" }]