A simple and convenient command-line interface (CLI) for interacting with the Context7 API.

Quickly search for projects, get information about them, and query specific topics within a project's context, all directly from your terminal. Provides the c7 command.

⚠️ Important: All data provided by this tool comes directly from the official Context7 API (https://context7.com). This tool is simply a convenient wrapper and does not generate or host any content itself.

• Query: Ask questions or request information on specific topics within a project's documentation context.
• Search: Discover available projects indexed by Context7 based on keywords.
• Info: Get metadata about a specific project (like its source repository, last update time, etc.).
• Flexible Output: Get query results in plain text (txt, default) or structured json.
• Save Results: Easily save query output directly to a file (llms_{project}.{format}).
• User-Friendly: Uses spinners for feedback during API calls and colored output for readability.

• Node.js (Version 18+ recommended for native fetch, but should work with v16+ if needed)
• npm (usually comes with Node.js) or Bun

npm install -g context7-cli

1. Clone the repository:

   git clone https://github.com/quiint/c7
   cd c7

2. Install dependencies:

   bun install

3. Run the CLI in the repository:

   bun run index.js

The basic command structure is:

c7 [command] [arguments...] [options...]

Fetches information about a specific topic within a project. This is the default command if no other command (search, info) is specified.

Syntax:

c7 <projectname> <query...> [options]

Arguments:

• <projectname>: The exact name of the project (e.g., nextjs, react, uv). Use c7 search <term> to find valid names. (Required)
• <query...>: The topic, question, or keywords to query within the project's context. (Required)

Options:

• --type, -t: Specify the output format.
  • txt (Default)
  • json
• --save, -s: Save the output to a file named llms_{projectname}.{format} in the current directory instead of printing to the console. (Boolean flag)
• --tokens, -k: Specify the maximum number of tokens for the API response. (e.g., --tokens 1000). The API default is used if omitted. (Number)

Examples:

# Get info about data fetching in Next.js (default TXT output)
c7 nextjs data fetching strategies

# Get info about state management in React, output as JSON
c7 react state management -t json

# Search for info on installation in the 'uv' project and save to llms_uv.txt
c7 uv installation steps --save

# Query Node.js streams and save the result as llms_nodejs.json
c7 nodejs streams api -t json -s

# Query Next.js data fetching and use only 1000 tokens
c7 nextjs data fetching -k 1000

Searches for available Context7 projects based on a keyword in their title. It filters out projects whose internal path contains a /, as these are often duplicates or not directly queryable via this tool.

Syntax:

c7 search <term>

Arguments:

• <term>: The keyword to search for in project titles (e.g., react, python, aws). (Required)

Example:

c7 search next
# Output might look like:
# --- Search Results (Use "Project Name" for queries) ---
#   - Next.js (Project Name: nextjs)
# -------------------------------------------------------
# Example query: c7 nextjs <your query>

Displays metadata associated with a specific project.

Syntax:

c7 info <projectname>

Arguments:

• <projectname>: The exact project name (e.g., nextjs, react). Use the name shown in parentheses from the search command results. (Required)

Example:

c7 info react
# Output will show details like Title, Project Name, Docs Source, Last Update, etc.

You can always get help on commands and options:

c7 --help          # General help for all commands
c7 query --help    # Help specifically for the query command (same as default)
c7 search --help   # Help for the search command
c7 info --help     # Help for the info command

1. Clone the repository.
2. Install dependencies (npm install or bun install).
3. Link for local testing (npm link or bun link).
4. Make your changes!

MIT