A spec-driven development workflow for AI coding assistants that agrees on proposals, specs, and tasks before any code is written.
Houseofmvps/codesight
A CLI that scans your codebase and generates compact AI context files so coding assistants spend fewer tokens exploring the project.
- Stars
- 1,419
- Forks
- 127
- Open issues
- 2
- Last commit
- 3 months ago
What is it?
codesight is a command-line tool that analyzes a project's routes, database models, UI components, libraries, environment variables, and import graph, then writes the results as structured markdown. Developers using AI coding assistants such as Claude Code, Cursor, GitHub Copilot, or Codex can load this single map instead of having the assistant grep and read dozens of files at the start of each conversation.
The tool runs with npx and needs no configuration or API keys. For TypeScript projects it uses the TypeScript compiler API to parse code structurally; other languages, including Python, Go, Ruby, Java, Rust, PHP, and Swift, use regex-based detection across the supported frameworks. It also includes a blast-radius command that lists files, routes, and models affected by changing a given file, an optional wiki mode that splits the context into per-topic articles, and an MCP server with 14 tools.
The project is written in TypeScript with zero runtime dependencies and targets Node.js 18 or later. Its README reports token-reduction benchmarks on several open-source and private projects.
Key features
- 01Generates route, schema, component, env var, and dependency maps as markdown files.
- 02Uses the TypeScript compiler API for precise parsing when TypeScript is installed.
- 03Supports Python, Go, Ruby, Java, Rust, PHP, Swift, Dart, and more via regex detection.
- 04Provides a blast-radius command to show files and routes affected by a change.
- 05Offers an MCP server with 14 tools for Claude Code and Cursor.
- 06Can generate CLAUDE.md, .cursorrules, codex.md, and AGENTS.md configuration files.
Use cases
- Reducing the context an AI coding assistant loads at the start of each session in a large monorepo.
- Checking which files, routes, and models a change to a shared module will affect before editing.
- Auditing required versus defaulted environment variables across a codebase.
- Giving a new team member or AI agent a quick structural overview of an unfamiliar project.
Topics
FAQ
- What is codesight?
- A CLI that scans your codebase and generates compact AI context files so coding assistants spend fewer tokens exploring the project. codesight is a command-line tool that analyzes a project's routes, database models, UI components, libraries, environment variables, and import graph, then writes the results as structured markdown. Developers using AI coding assistants such as Claude Code, Cursor, GitHub Copilot, or Codex can load this single map instead of having the assistant grep and read dozens of files at the start of each conversation.
- Is codesight free to use? What license does it have?
- codesight is open source under the MIT license. Check the license terms before using it in commercial products.
- Is codesight still maintained?
- Updated within 6 months. The last commit to codesight was on Jul 27, 2026.
- What can I use codesight for?
- Reducing the context an AI coding assistant loads at the start of each session in a large monorepo. Checking which files, routes, and models a change to a shared module will affect before editing. Auditing required versus defaulted environment variables across a codebase. Giving a new team member or AI agent a quick structural overview of an unfamiliar project.
- Where can I get codesight?
- The source code and installation instructions are on GitHub: https://github.com/Houseofmvps/codesight
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An open-source AI coding agent that reads your project, edits files and runs commands and tests from your terminal, desktop, browser or editor.
An MCP server that gives coding agents IDE-level symbol search, semantic editing and refactoring over language servers or JetBrains IDEs.
A local code knowledge graph that gives AI coding tools precise review context over MCP and a CLI.
An agentic coding tool that runs in your terminal, reads your codebase, and handles coding tasks, code explanations, and git workflows through natural language.
A Python framework and runtime for building agent platforms, serving agents as an API, and managing them through a web UI.