ai-bridge-cli

Token-budget aware dependency crawler for assembling AI chat context.

Install

npm install -g ai-bridge-cli

Usage

# Crawl a file and its dependencies, copy to clipboard
bridge-context src/pages/Dashboard.tsx

# Set a custom token budget and depth
bridge-context src/App.tsx --budget 10000 --depth 3

# Skip interactive mode, output to file
bridge-context src/index.ts --no-interactive --output context.md

# Include database schema
bridge-context src/api/routes.ts --db postgres

Options

| Flag | Default | Description | |------|---------|-------------| | -b, --budget <tokens> | 8000 | Token budget (stops when exceeded) | | -d, --depth <N> | 2 | Maximum crawl depth for imports | | --include-libs | false | Include library import names | | --no-interactive | - | Skip import selection checklist | | --db <connector> | - | Include external context (e.g., DB schema) | | --exclude <patterns> | node_modules,dist,... | Comma-separated exclude patterns | | --no-copy | - | Don't copy to clipboard | | -o, --output <file> | - | Write to file instead of clipboard |

How It Works

1. Parses the entry file for import / require statements
2. Shows an interactive checklist of local dependencies
3. Crawls selected imports via BFS up to --depth
4. Counts tokens using tiktoken (cl100k_base encoding)
5. Stops when the token budget is exceeded
6. Formats output as Markdown with code blocks
7. Copies to clipboard (ready to paste into AI chat)

Database Connectors

Create .bridge-context.json in your project root:

{
  "dbConnectors": {
    "postgres": "node scripts/dump_postgres_schema.js --table users",
    "redis": "node scripts/dump_redis_keys.js"
  }
}

Then use --db postgres to include the output.

License

MIT