ChronicleMCP

Table of Contents

1. Quick Start
2. Features
3. Installation
4. Usage
   • CLI Commands
   • MCP Tools
   • HTTP API
5. Configuration
6. Security & Privacy
7. Supported Browsers
8. Troubleshooting
9. Development
10. Contributing
11. License

Quick Start

# Install
pip install chronicle-mcp

# Run MCP server (stdio mode for AI agents like Claude, Cursor)
chronicle-mcp mcp

# Or start an HTTP server
chronicle-mcp http --port 8080

# Check available browsers
chronicle-mcp list-browsers

Features

Installation

pip (Recommended)

pip install chronicle-mcp

pipx (Isolated Installation)

pipx install chronicle-mcp

Docker

# Pull the latest image
docker pull ghcr.io/nikolasil/chronicle-mcp:latest

# Run the server
docker run -p 8080:8080 ghcr.io/nikolasil/chronicle-mcp

From Source

git clone https://github.com/nikolasil/chronicle-mcp.git
cd chronicle-mcp
pip install -e .

Homebrew (macOS)

brew tap nikolasil/chronicle-mcp
brew install chronicle-mcp

Usage

CLI Commands

chronicle-mcp mcp

Run the MCP server for AI agents.

# Stdio mode (default for AI assistants)
chronicle-mcp mcp

# SSE mode for HTTP clients
chronicle-mcp mcp --sse --port 8080

chronicle-mcp http

Start a long-running HTTP REST API server.

# Default (foreground, port 8080)
chronicle-mcp http

# Custom port
chronicle-mcp http --port 9000

# Daemon mode
chronicle-mcp http --port 8080 --daemon

# Different browser
chronicle-mcp http --browser firefox

chronicle-mcp status

Check if the server is running.

chronicle-mcp status --port 8080

chronicle-mcp logs

View server logs.

chronicle-mcp logs --port 8080 --lines 100

chronicle-mcp version

Show version information.

chronicle-mcp version

chronicle-mcp list-browsers

List available browsers on the system.

chronicle-mcp list-browsers
# Output: Available browsers: chrome, edge

chronicle-mcp completion

Generate shell completion scripts.

# Bash
chronicle-mcp completion bash >> ~/.bashrc

# Zsh
chronicle-mcp completion zsh >> ~/.zshrc

# Fish
chronicle-mcp completion fish > ~/.config/fish/completions/chronicle-mcp.fish

MCP Tools

search_history

Search browser history for keywords in titles or URLs.

def search_history(
    query: str,
    limit: int = 5,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

# Markdown output (default)
search_history("python tutorial", limit=10, browser="chrome")

# JSON output
search_history("github", limit=5, format_type="json")

get_recent_history

Get recent browsing history from the last N hours.

def get_recent_history(
    hours: int = 24,
    limit: int = 20,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

# Last 24 hours
get_recent_history(hours=48, limit=20)

count_visits

Count total visits to a specific domain.

def count_visits(
    domain: str,
    browser: str = "chrome"
) -> str

Example:

count_visits("github.com", browser="chrome")
# Output: Visits to 'github.com' in chrome: 42

list_top_domains

Get the most visited domains.

def list_top_domains(
    limit: int = 10,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

list_top_domains(limit=20)

search_history_by_date

Search history within a date range.

def search_history_by_date(
    query: str,
    start_date: str,  # ISO format: YYYY-MM-DD
    end_date: str,    # ISO format: YYYY-MM-DD
    limit: int = 10,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

search_history_by_date(
    "python",
    start_date="2024-01-01",
    end_date="2024-12-31",
    limit=20
)

list_available_browsers

Returns a list of browsers with detected history databases.

def list_available_browsers() -> str

Example:

list_available_browsers()
# Output: Available browsers: chrome, edge, firefox, brave

delete_history

Delete history entries matching a query.

def delete_history(
    query: str,
    limit: int = 100,
    browser: str = "chrome",
    confirm: bool = False
) -> str

Example:

# Preview what would be deleted (default)
delete_history("spam.com", browser="chrome")
# Actually delete
delete_history("spam.com", browser="chrome", confirm=True)

search_by_domain

Search history within specific domain(s).

def search_by_domain(
    domain: str,
    query: str = None,
    limit: int = 20,
    browser: str = "chrome",
    format_type: str = "markdown",
    exclude_domains: list[str] = None
) -> str

Example:

search_by_domain("github.com", browser="chrome")

get_browser_stats

Get browsing statistics for the browser database.

def get_browser_stats(browser: str = "chrome") -> str

Example:

get_browser_stats(browser="firefox")
# Returns JSON with total visits, domains, date range, etc.

get_most_visited_pages

Get the most visited individual pages.

def get_most_visited_pages(
    limit: int = 20,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

get_most_visited_pages(limit=10, browser="chrome")

export_history

Export history to CSV or JSON format.

def export_history(
    format_type: str = "csv",
    limit: int = 1000,
    query: str = None,
    browser: str = "chrome"
) -> str

Example:

# Export to CSV
export_history(format_type="csv", limit=1000, browser="chrome")

# Export to JSON
export_history(format_type="json", limit=500, browser="firefox")

search_history_advanced

Advanced search with multiple options including regex and fuzzy matching.

def search_history_advanced(
    query: str,
    limit: int = 20,
    browser: str = "chrome",
    format_type: str = "markdown",
    exclude_domains: list[str] = None,
    sort_by: str = "date",
    use_regex: bool = False,
    use_fuzzy: bool = False,
    fuzzy_threshold: float = 0.6
) -> str

Example:

# Fuzzy search
search_history_advanced("pythn tutorial", use_fuzzy=True, fuzzy_threshold=0.7)

# Regex search
search_history_advanced(r"github\.com/.*/issues/\d+", use_regex=True)

sync_history

Sync history between browsers.

def sync_history(
    source_browser: str,
    target_browser: str,
    merge_strategy: str = "latest",
    dry_run: bool = True
) -> str

Example:

# Preview sync
sync_history("chrome", "firefox", dry_run=True)

# Actually sync
sync_history("chrome", "firefox", dry_run=False)

list_available_bookmarks

Returns a list of browsers with detected bookmarks.

def list_available_bookmarks() -> str

Example:

list_available_bookmarks()
# Output: Available browsers with bookmarks: chrome, edge, firefox

list_available_downloads

Returns a list of browsers with detected downloads history.

def list_available_downloads() -> str

Example:

list_available_downloads()
# Output: Available browsers with downloads: chrome, edge, firefox

get_bookmarks

Get bookmarks from a browser.

def get_bookmarks(
    query: str = None,
    limit: int = 50,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

# Get all bookmarks
get_bookmarks(browser="chrome")

# Search bookmarks
get_bookmarks(query="python", browser="firefox")

get_downloads

Get downloads history from a browser.

def get_downloads(
    query: str = None,
    limit: int = 50,
    browser: str = "chrome",
    format_type: str = "markdown"
) -> str

Example:

# Get all downloads
get_downloads(browser="chrome")

# Search downloads
get_downloads(query="pdf", browser="firefox")

HTTP API

Endpoints

Request/Response Examples

POST /api/search

curl -X POST http://localhost:8080/api/search \
  -H "Content-Type: application/json" \
  -d '{"query": "python tutorial", "limit": 10, "browser": "chrome", "format": "markdown"}'

Response:

{
  "results": [
    {
      "title": "Python Tutorial",
      "url": "https://docs.python.org/3/tutorial/",
      "timestamp": "2024-01-15T10:30:00+00:00"
    }
  ],
  "count": 1
}

GET /health

curl http://localhost:8080/health

Response:

{
  "status": "healthy",
  "service": "chronicle-mcp",
  "version": "1.4.0",
  "timestamp": "2024-01-15T10:30:00+00:00"
}

Configuration

ChronicleMCP can be configured using environment variables or a config file.

Environment Variables

Config File

Create ~/.config/chronicle-mcp/config.toml:

[default]
browser = "chrome"
limit = 10
format = "markdown"
log_level = "INFO"

Security & Privacy

• Local-Only: All data stays on your machine
• URL Sanitization: Sensitive query parameters are automatically removed:
  • token, session, key, password, auth, sid, access_token
• Temp Files: History is copied to temporary files that are cleaned up after each query
• No Data Collection: Your browsing data is never sent to any server
• Error Messages: No sensitive file paths are exposed

Supported Browsers

ChronicleMCP supports reading history from the following browsers:

Troubleshooting

"Browser history not found"

• Ensure the browser is installed
• Check that Chrome/Edge isn't currently open (locks the database)
• Run chronicle-mcp list-browsers to see detected browsers

"Permission denied"

• Check file permissions on the browser's history database
• On Windows, ensure the browser is closed before querying

Empty results

• Try a more specific search term
• Check the date range for search_history_by_date
• Verify the browser has history data

Performance issues

• Large history databases may take longer to query
• Consider reducing the limit parameter

Development

Running Tests

# Run all tests
pytest

# Verbose output
pytest -v

# Specific test file
pytest tests/unit/

# Single test
pytest -k test_name

# With coverage
pytest --cov=chronicle_mcp

Development Server

# Run MCP server (stdio mode)
python -m chronicle_mcp mcp

# Run HTTP server
python -m chronicle_mcp http --port 8080

Code Quality

# Linting
ruff check .

# Formatting
ruff format .

# Type checking
mypy chronicle_mcp/

Contributing

Contributions are welcome! Please read our Contributing Guide for details.

License

MIT License - See LICENSE file for details.