Skip to content
codectx

Quick Start

The fastest way to understand what codectx does is to run it on your own codebase.

  1. Navigate to your repository Open your terminal and navigate to any Python, JavaScript, TypeScript, or Rust project.

    Terminal window
    cd ~/my-awesome-repo

  2. Run codectx Run the analyze command specifying the current directory (.) as the target:

    Terminal window
    codectx analyze .

  3. Check the Output codectx will generate a structured CONTEXT.md file in the root of your project. Open it in a text editor to see a tiered layout detailing your architectural entry points, dependency connections, and key implementations.

  4. Try a specific task profile codectx can adjust its scoring and output based on what your AI assistant is trying to accomplish. For example, if you are asking an agent to write a new feature, run:

    Terminal window
    codectx analyze . --task feature

Passing Context to an LLM

Section titled “Passing Context to an LLM”

Now that you have your CONTEXT.md, here is how you use it:

  1. Open your AI coding assistant (Claude, cursor, ChatGPT, etc.).
  2. Drag and drop the CONTEXT.md file into the chat.
  3. Prompt the agent: “Read the attached context file to understand my project’s architecture, then implement [Feature X] in the [Module Y].”

By providing the structured CONTEXT beforehand, the LLM starts with a deterministic, highly organized understanding of exactly how your project goes together.

Next, learn more commands in Basic Usage.