Jinni: Bring Your Project Into Context

What is Jinni: Bring Your Project Into Context?

Jinni is a tool for developers to efficiently provide Large Language Models with a consolidated view of relevant project files. It integrates with AI tools via an MCP (Model Context Protocol) server and offers a command-line utility for manual copy-paste or file output.

How to use Jinni: Bring Your Project Into Context?

Configure your MCP client to run the server via uvx jinni-server, or use the CLI directly with jinni [OPTIONS] [<PATH...>]. The MCP server exposes the read_context tool; the CLI outputs concatenated file contents to stdout or clipboard.

Key features of Jinni: Bring Your Project Into Context

• Efficient consolidation of relevant project files.
• Intelligent filtering using gitignore-style patterns.
• Customizable inclusions via .contextfiles.
• Configurable size limit for large contexts.
• File path headers in output.
• List-only mode for file preview.

Use cases of Jinni: Bring Your Project Into Context

• Integrating with AI coding assistants like Cursor or Claude Desktop.
• Quickly providing project context to an LLM for code generation.
• Focusing LLM attention on specific modules or file patterns.
• Debugging large projects by excluding irrelevant directories.
• Generating a structured text dump for offline LLM analysis.

FAQ from Jinni: Bring Your Project Into Context

What runtime does Jinni require?

Jinni requires uv to run the MCP server via uvx jinni-server. The CLI is also available as jinni and can be run directly.

How does Jinni decide which files to include?

It uses built-in defaults that exclude binaries, dotfiles, and common temp/build directories. It also respects .gitignore rules and allows custom inclusions via .contextfiles