# Identity Server

A local MCP server that maintains and serves personal identity context to LLMs.

## Installation

```bash
# Clone and set up virtual environment
git clone https://github.com/ruthvik-947/identity-server.git
cd identity-server
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Install dependencies
pip install -e .
```

> **Note:** Always run commands from the virtual environment. The project uses sentence-transformers and sqlite-vec for semantic search, which require specific dependency versions.

## Quick Start

1. Initialize your identity directory:

```bash
identity init
```

2. Edit `~/.identity/identity.yaml` with your info

3. Configure sources in `~/.identity/sources.yaml`

4. Sync your sources:

```bash
identity sync
```

5. Add to Claude Code's MCP config:

```json
{
  "mcpServers": {
    "identity": {
      "command": "/path/to/identity-server/.venv/bin/identity",
      "args": ["serve"]
    }
  }
}
```

> Replace `/path/to/identity-server` with your actual installation path (e.g., `~/.local/identity-server`).

## CLI Commands

- `identity init` - Initialize identity directory
- `identity status` - Show current identity
- `identity sync` - Sync all sources
- `identity sources` - List configured sources
- `identity serve` - Start MCP server

## MCP Tools

- `get_identity()` - Core identity info
- `get_projects()` - Indexed projects
- `get_reading()` - Reading list items
- `query(question)` - Semantic search
- `list_sources()` - Configured sources
- `add_source(description)` - Setup instructions

## Source Types

- `local_files` - Markdown/text files via glob
- `git_repos` - README, commits, languages
- `rss` - RSS/Atom feeds
- `custom` - Python script connectors

## Custom Connectors

Create Python scripts to pull data from any source (APIs, databases, etc.).

**1. Create a connector in `~/.identity/connectors/`:**

```python
# ~/.identity/connectors/my_connector.py

def fetch(config):
    """
    Args:
        config: dict from sources.yaml 'config' field

    Returns:
        list of dicts with: title, body, tags (optional), timestamp (optional)
    """
    return [
        {
            "title": "Example Item",
            "body": "Content goes here...",
            "tags": ["tag1", "tag2"],
        }
    ]
```

**2. Add to `sources.yaml`:**

```yaml
sources:
  - name: my-source
    type: custom
    connector: connectors/my_connector.py  # relative to ~/.identity
    privacy: private
    config:  # passed to fetch()
      api_key: "..."
```

**3. Run `identity sync` to execute connectors and index results.**

Connectors only run during sync—the MCP server queries the local database.

## Privacy

Tag sources as `public` or `private`. Private content is excluded from MCP queries.