CodeRabbit Plan turns your ideas into detailed, project-aware Coding Plans you can review, refine, and hand off to any coding agent. Because every plan is grounded in deep codebase understanding and enriched by your project Knowledge Base, it references the right files, follows your established patterns, and produces agent-ready prompts, not generic outlines.Documentation Index
Fetch the complete documentation index at: https://docs.coderabbit.ai/llms.txt
Use this file to discover all available pages before exploring further.
1. Create a plan
- Using web interface
- Using issue tracker
The fastest way to get started is the CodeRabbit web app. Go to app.coderabbit.ai/plan, describe what you want to build in a free-form text, select a repository, and click Create plan. The form also supports file attachments.See the Create a plan guide for the full step-by-step walkthrough.
Codebase understanding
Analyzes your repository’s architecture, patterns, and existing code so
every plan fits naturally into how your project is built.
Project context
Draws on related issues, design documents, and accumulated to
ground every plan in your project’s broader context.
2. Refine
Chat with CodeRabbit to refine details, challenge design choices, or request changes. The conversation is collaborative: product owners and other team members can participate too, ensuring the plan reflects the team’s knowledge.- Web app
- Issue Trackers (GitHub & GitLab)
Review, tune, and adjust your Coding Plan in the CodeRabbit web app. Use the
chat panel to iterate on the plan until it’s ready.

Collaborative Planning
Plans are available for review by engineers and product owners. Team members discuss, challenge design choices, and refine plans together.
Accountability and history
Every plan version is preserved. Track what was planned, when it was planned, and why decisions were made.
3. Handoff
Once the plan reflects the team’s decisions, hand off the finalized prompts to your coding agent of choice. The agent receives precise, codebase-aware instructions it can act on immediately.Agent-ready prompts
Finalized plans are exported as structured, codebase-aware prompts that any coding agent can act on immediately — no reinterpretation required.
Any coding agent
Hand off to Claude Code, Codex, Cursor, Gemini, or any other coding agent. The prompt format is agent-agnostic.
What’s next
CodeRabbit Plan documentation
Full documentation covering the web interface, plan structure, refinement, and Agent Handoff.
Issue-based planning
Platform-specific guides for generating Coding Plans from GitHub, GitLab, Jira, and Linear issues.
CodeRabbit Agent for Slack
Bring AI-powered investigation, planning, and code editing into your Slack workspace.
