Metadata-Version: 2.4
Name: gutfeel-mcp
Version: 0.7.4
Summary: gut's judgment calls as MCP tools: likely, classify, rate and each, answered by small models.
Keywords: mcp,model-context-protocol,classification,jev,llm,agents
Author: Eren Kahraman
Author-email: Eren Kahraman <kahraman1234eren@gmail.com>
License-Expression: Apache-2.0
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Dist: gutfeel==0.7.4
Requires-Dist: mcp>=2.2,<3
Requires-Dist: typesafe-sdk>=0.7
Requires-Python: >=3.10
Project-URL: Homepage, https://gutpy.dev/docs/mcp.html
Project-URL: Repository, https://github.com/Kungie/gut
Description-Content-Type: text/markdown

# gutfeel-mcp

gut's judgment calls as MCP tools, for Claude Code, Claude Desktop, Cursor and any other MCP client:
`likely`, `classify`, `rate` and `each`, answered by a small, fast, cheap model, with YES, NO or
UNSURE.

```bash
claude mcp add gut --env TYPESAFE_API_KEY=your-key -- uvx gutfeel-mcp
```

Or, in any client's JSON configuration:

```json
{
  "mcpServers": {
    "gut": {
      "command": "uvx",
      "args": ["gutfeel-mcp"],
      "env": { "TYPESAFE_API_KEY": "your-key" }
    }
  }
}
```

The model comes from the environment: TypeSafe's Jev whenever `TYPESAFE_API_KEY` is set, or
`GUT_BACKEND` for another -- `openrouter`, `ollaya`, `ollama`, `openai`, or `zeroshot` for a local
model (`uvx --with "gutfeel[local]" gutfeel-mcp`).

Documentation: https://gutpy.dev/docs/mcp.html · The library: https://pypi.org/project/gutfeel/

<!-- mcp-name: io.github.Kungie/gut -->
