Metadata-Version: 2.4
Name: mcp-server-flockx
Version: 0.2.0
Summary: MCP server for FlockX AI agent platform - enables AI assistants to interact with FlockX agents
Author-email: FlockX Team <support@flockx.io>, Mark Alosey <mark@flockx.io>
License: MIT
Project-URL: Homepage, https://flockx.io
Project-URL: Source, https://github.com/flockx-official/mcp-server-flockx
Project-URL: Bug Reports, https://github.com/flockx-official/mcp-server-flockx/issues
Project-URL: Documentation, https://docs.flockx.io
Keywords: mcp,flockx,ai,agents,cursor,claude
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: requests<3,>=2.32
Requires-Dist: pydantic<3,>=2.8
Requires-Dist: pydantic-settings<3,>=2.3
Requires-Dist: fastmcp<3,>=2.7.1
Requires-Dist: fastapi<1.0,>=0.115
Requires-Dist: uvicorn<1.0,>=0.30
Requires-Dist: python-dotenv<2,>=1.0
Requires-Dist: httpx<0.29,>=0.28.1
Dynamic: license-file

<!-- FlockX Logo: size matches header text -->
<p align="center">
  <img src="static/flockx-purple-logo.png" alt="FlockX Logo" style="height: 2em; vertical-align: middle;" />
  <br/>
  <span style="font-size: 1em; font-style: italic; color: #555;">FlockX MCP Server</span>
</p>

# FlockX MCP Server

[![PyPI version](https://badge.fury.io/py/mcp-server-flockx.svg)](https://badge.fury.io/py/mcp-server-flockx)
[![Python](https://img.shields.io/pypi/pyversions/mcp-server-flockx.svg)](https://pypi.org/project/mcp-server-flockx/)

> **Legendary AI Agent Creation & Knowledge Ingestion for Cursor and Claude Desktop**
> and so, so, so much more

---

## 🚀 Overview

This repository provides a Model Context Protocol (MCP) server for FlockX, enabling you to manage, train, and validate AI agents through a seamless conversational interface—right inside your development environment (Cursor) or Claude Desktop. With simple Markdown processes and best-practice templates, you can automate agent creation, knowledge ingestion, and validation with ease.

---

## 🛠️ Installation & Setup

### Prerequisites

- Python 3.8+
- FlockX API key (get one at [agents.flockx.io](https://agents.flockx.io))

### Quick Install (Recommended)

```bash
# Install from PyPI
pip install mcp-server-flockx

# Or use uvx for isolated execution (preferred for MCP)
uvx mcp-server-flockx

# Or use pipx for global installation
pipx install mcp-server-flockx
```

### Alternative: Install from Source

```bash
git clone https://github.com/flockx-official/mcp-server-flockx
cd mcp-server-flockx
pip install -e .
```

---

## 🖥️ Integrate with Cursor IDE

1. Open Cursor Settings (Cmd+, or Ctrl+,)
2. Search for "MCP" or navigate to MCP settings
3. Add to your MCP configuration:

```json
{
  "mcpServers": {
    "flockx-mcp": {
      "command": "uvx",
      "args": ["mcp-server-flockx"],
      "env": {
        "FLOCKX_API_KEY": "your_api_key_here"
      }
    }
  }
}
```

4. Restart Cursor
5. **Enable Agent Mode** in your chat (critical!)
6. Your FlockX MCP tools will appear in the tools panel

---

## 💬 Integrate with Claude Desktop

1. Open Claude Desktop → Settings → Developer → Edit Config
2. Add to `claude_desktop_config.json`:

```json
{
  "mcpServers": {
    "flockx-mcp": {
      "command": "uvx",
      "args": ["mcp-server-flockx"],
      "env": {
        "FLOCKX_API_KEY": "your_api_key_here"
      }
    }
  }
}
```

3. Save and restart Claude Desktop
4. Look for the MCP tools icon in chat

---

## 🤖 Why Use the flockx MCP Server?

Working with the flockx MCP server in Cursor or Claude Desktop unlocks a new paradigm:

- **Conversational Training:** Interact with your agent, teach new skills, and ingest knowledge—all through natural language.
- **Process Automation:** Define and run complex workflows using simple Markdown files in `training/processes/`. No code required—just describe your process, and let your AI do the rest.
- **Best Practice Templates:** Use the provided templates in `training/templates/` to ensure every knowledge document and agent profile is structured for success.

You can iteratively build, test, and refine your agents in a collaborative, conversational loop—directly in your favorite environment.

## 🧪 What This Workflow Does For You

When you follow this workflow, the system will:

- **Create test agents in your FlockX account** automatically.
- **Train each agent on your provided profession or domain knowledge** (using your input files and best-practice extraction).
- **Enable you to ask any of your agents questions**—directly in Cursor or Claude Desktop—so you can interact, test, and validate their knowledge in real time.

This makes it easy to prototype, experiment, and iterate on agent skills and knowledge before deploying to production.

---

## 🚀 Full Workflow Overview

See how much gets done with a single command:

```mermaid
flowchart TD
  subgraph "Extract Skills Knowledge Local"
    A[List Inputs] --> B[Read Input File]
    B --> C{Extract Skills}
    C --> D[Fill Knowledge Template]
    D --> E[Save Knowledge Doc]
  end

  subgraph "Generate Agent Profiles Local"
    F[List Knowledge Docs] --> G[Read Knowledge Doc]
    G --> H{Extract Agent Info}
    H --> I[Fill Agent Template]
    I --> J[Save Agent Profile]
  end

  subgraph "Create Agents FlockX API"
    K[Create Agent API]
  end

  subgraph "Upload Knowledge FlockX API"
    L[Upload Knowledge Doc API]
  end

  subgraph "Validate Completion FlockX API"
    M[Query FlockX API]
  end

  E --> F
  J --> K
  E --> L
  K --> M
  L --> M

  classDef api_node fill:#e0e7ff,stroke:#6366f1,stroke-width:2px;
  class K,L,M api_node
```

---

## 📝 Next Steps: Orchestrate the Full Workflow

To get started with end-to-end agent creation and knowledge ingestion, simply ask your AI assistant (in Cursor or Claude Desktop):

> **"Please follow the instructions in `training/processes/process-overall-agent-knowledge-ingestion.md` to extract, generate, upload, and validate agent knowledge."**

Your AI will guide you through each step, leveraging the Markdown processes and templates provided.

---

## 🛠️ Troubleshooting

- **`FLOCKX_API_KEY` not set:** Add your API key to the MCP configuration in Cursor/Claude settings
- **MCP server not responding:** Ensure you've restarted your IDE after configuration
- **"Need approval" messages:** Enable Agent Mode in Cursor (critical!)
- **Installation issues:** Try `pip install --upgrade mcp-server-flockx`
- **Rate limiting:** If you encounter API rate limits, consider adjusting your workflow or contacting FlockX support

---

## 📦 Releasing to PyPI (maintainers)

Semantic versioning is used: MAJOR.MINOR.PATCH.

1. Update `pyproject.toml` and `src/mcp_server_flockx/__init__.py` with the new version.
2. Update `CHANGELOG.md`.
3. Create and push a tag to trigger publish (GitHub Actions):

```bash
git checkout main && git pull
git tag -a v0.2.0 -m "Release v0.2.0"
git push origin v0.2.0
```

CI will build and publish to PyPI when the tag is pushed. Configure repository secret `PYPI_API_TOKEN` with an API token from PyPI.

## 🤝 Contributing
