Gread

English

中文

English

中文

Appearance

GreadOne Skill for all repos

Give your agent access to the source code and docs of all open-source libraries

Quick Start

Plug and Play

Compatible with all popular coding agents and MCP clients, ready to use out of the box

One Skill, All Repos

Can access all public GitHub repositories

Skill and MCP

Both Skill and MCP are provided, connect with either option

Docs Integration

Automatically recognizes docs repo and provides them along with the main repo when accessed

Efficient Access

Provides multiple tools and parameters, responds quickly after the first indexing

100% Free & Open Source

We offer free services publicly and are open-sourced under Apache 2.0

Quick Start ​

We offer two integration methods, choose either one:

  • Skill: Suitable for coding agents like OpenCode, Codex, Cursor, Copilot, etc.
  • MCP: Suitable for AI chat apps and other MCP clients

Skill ​

Run the following command to install:

bash
npx skills add https://github.com/NitroRCr/gread --skill gread

MCP ​

Streamable HTTP URL:

https://api.gread.dev/mcp

JSON configuration reference:

json
{
  "mcpServers": {
    "gread": {
      "type": "streamableHttp",
      "url": "https://api.gread.dev/mcp"
    }
  }
}

More Information ​

You can view SKILL.md to understand the tools and parameters provided by Gread. Or, directly add Gread and ask your Agent for more information!

Comparison ​

Understand how Gread differs from existing tools:

Compare with Deepwiki / Zread
  • Deepwiki / Zread can only ask AI on their websites, while Gread brings this capability to your own Agent
  • Deepwiki / Zread's AI can only access the current repo and cannot access multiple repos simultaneously; Gread can access any repo at any time
  • Deepwiki / Zread are closed-source, while Gread is open-source
  • Deepwiki / Zread generates documentation for in-depth project understanding, while Gread does not have this functionality
Compare with Zread MCP
  • Zread MCP requires purchasing a GLM Coding Plan to use and has strict usage limits; Gread is free to use
  • Zread MCP is closed-source, while Gread is open-source
  • Zread MCP only provides MCP, while Gread provides both Skill and MCP
Compare with Context7
  • Context7 mainly focuses on docs; whereas Gread mainly focuses on source code, while also being able to access docs in the form of source code
  • Context7 only open-sourced its MCP server, its backend is proprietary; whereas Gread is completely open-source
  • Context7 has relatively strict usage limits and paid plans; whereas Gread is completely free
  • Gread can access all public GitHub repositories; Context7 can only access submitted libraries, but supports more sources
Compare with GitMCP
  • GitMCP uses GitHub API to read/search code, with slower response times; Gread clones the code to the server locally and uses git to read/search code, with rapid response.
  • GitMCP does not provide a tool to list directory trees, while Gread does. This helps the Agent read codebases more efficiently.
  • GitMCP only provides MCP, while Gread provides both Skill and MCP
Compare with Sourcebot
  • Sourcebot can only access indexed repositories; Gread can access all public repositories, "indexing" is just a cache
  • Sourcebot requires self-deployment, while Gread provides out-of-the-box services
  • Sourcebot provides a user interface and MCP, while Gread provides Skill and MCP

Online Experience ​

Want to try Gread online? Check out Nya AI. It has Gread built-in—simply enable the "GitHub Repos" plugin in the chat.