Skip to main content

Local-first LLM model deprecation watchdog — scans your configs, alerts on sunsets, and instructs your IDE to update.

Project description

Chowkidar

(Dependabot for your LLMs)

PyPI Version PyPI Downloads VS Code Marketplace Version VS Code Marketplace Downloads

Local-first LLM model deprecation watchdog.

Chowkidar scans your project configs for LLM model strings, cross-references them against a local deprecation registry, and alerts you before models sunset — via desktop notifications and IDE rules that instruct AI assistants to update deprecated models automatically.

Everything runs on your machine. Zero data exfiltration.

Quick Start

pip install chowkidar

# First-time setup (initializes config + database)
chowkidar setup --skip-slm

# Fetch deprecation data from providers
chowkidar sync

# Scan your project
chowkidar scan .

# Check for deprecated models
chowkidar check .

Features

  • Multi-format scanning: .env, YAML, TOML, JSON, Python, JavaScript, TypeScript
  • Provider coverage: OpenAI, Anthropic, Google, Mistral (extensible plugin architecture)
  • IDE rules (zero-config): Auto-generates rules files for Cursor, Claude Code, VS Code/Copilot, Windsurf
  • MCP server: Interactive tools for querying deprecation status from your IDE
  • Desktop notifications: Threshold-based alerts (90d, 30d, 7d, sunset)
  • Background daemon: Periodic scanning with OS-native service installation
  • Local SLM: Optional Ollama integration for parsing unstructured deprecation announcements
  • Safe updates: File locking, atomic writes, automatic backups, dry-run mode
  • Cross-platform: macOS, Linux, Windows

Commands

chowkidar setup [--skip-slm]     # Initialize config, DB, and optional SLM
chowkidar scan [PATH]            # Scan for model strings
chowkidar sync                   # Fetch deprecation data
chowkidar check [PATH]           # Check for deprecated models
chowkidar status                 # Show daemon status and watched projects
chowkidar watch <PATH>           # Register project for monitoring
chowkidar unwatch <PATH>         # Unregister project
chowkidar pin <MODEL> [--reason] # Suppress alerts for a model
chowkidar unpin <MODEL>          # Re-enable alerts
chowkidar snooze <MODEL> --days  # Temporarily suppress alerts
chowkidar daemon                 # Start background daemon
chowkidar install-service        # Install OS-native service
chowkidar mcp                    # Start MCP server (for IDE)
chowkidar config [KEY] [VALUE]   # View/set configuration
chowkidar update [--dry-run]     # Update deprecated models in .env
chowkidar rules write [PATH]     # Generate IDE rules files
chowkidar rules clean [PATH]     # Remove generated rules files
chowkidar slm status             # Check SLM availability

IDE Integration

Automatic Rules (Recommended)

Chowkidar writes rules files that AI assistants auto-discover — no configuration needed. If standard desktop notifications about model deprecation are ignored or snoozed, Chowkidar acts as your ultimate fallback: it auto-updates your editor's rules to ensure your AI model knows to update the deprecated model automatically.

Editor Rules File
Cursor .cursor/rules/chowkidar-alerts.mdc
Claude Code .claude/rules/chowkidar-alerts.md
VS Code/Copilot .github/copilot-instructions.md
Windsurf .windsurfrules

Run chowkidar rules write or let the daemon do it automatically.

MCP Server (Advanced)

Add to your IDE's MCP config:

{
  "mcpServers": {
    "chowkidar": {
      "command": "chowkidar",
      "args": ["mcp"]
    }
  }
}

Configuration

Config file: ~/.chowkidar/config.toml

Key Default Description
auto_update false Allow automatic .env modifications
write_rules true Generate IDE rules files
gitignore_rules true Add rules files to .gitignore
slm_enabled false Use local SLM for parsing
slm_model gemma3:1b Ollama model for SLM
scan_interval_hours 4 How often to scan watched projects
sync_interval_hours 24 How often to fetch provider data

Security

  • Zero exfiltration: No env content, API keys, or paths leave your machine
  • Read-only by default: File modification requires explicit auto_update = true
  • Atomic writes: All modifications use temp file + os.replace
  • Automatic backups: .env.chowkidar.bak created before any change
  • File locking: Prevents concurrent write corruption

License

MIT

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

chowkidar-0.3.0.tar.gz (724.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

chowkidar-0.3.0-py3-none-any.whl (69.9 kB view details)

Uploaded Python 3

File details

Details for the file chowkidar-0.3.0.tar.gz.

File metadata

  • Download URL: chowkidar-0.3.0.tar.gz
  • Upload date:
  • Size: 724.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.10

File hashes

Hashes for chowkidar-0.3.0.tar.gz
Algorithm Hash digest
SHA256 221be0dccaf6dadae5f01467da3539dc3e2b6e3a36d05d6b24fc0fdc17cfcbbb
MD5 2fe557e97c8b41bd8270b463edfb6f1a
BLAKE2b-256 f181cdae15090cbd3148450b4b2bb83abcb4dc9dc23a1dc7a9383c53e876cc08

See more details on using hashes here.

File details

Details for the file chowkidar-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: chowkidar-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 69.9 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.10

File hashes

Hashes for chowkidar-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 57ce8f2f8806fa7bc7c703ef9df9e55437f36d361c00df58c76b4a59100b8cc6
MD5 7ca769c067e49cd2fd49d46d326e5edf
BLAKE2b-256 5834a6ca79625315683ba93db4b4c268edf9a9bdf072919d803db3f24c4c8c2c

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page