# Antigravity — Repo Q&A Engine + Optional MCP

> A multi-agent repo knowledge engine (`ag-refresh`, `ag-ask`) that makes a codebase queryable, with optional MCP; verified 1216★, pushed 2026-05-14.

## Install

Merge the JSON below into your `.mcp.json`:

## Quick Use

```bash
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=cli"
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=engine"
cat > .env <<EOF
OPENAI_BASE_URL=https://your-endpoint/v1
OPENAI_API_KEY=your-key
OPENAI_MODEL=your-model
AG_ASK_TIMEOUT_SECONDS=120
EOF
ag-refresh --workspace .
ag-ask "How does auth work in this project?"
```

## Intro

A multi-agent repo knowledge engine (`ag-refresh`, `ag-ask`) that makes a codebase queryable, with optional MCP; verified 1216★, pushed 2026-05-14.

**Best for:** Teams onboarding large repos who want evidence-rich answers with file paths and routing

**Works with:** Claude Code, Codex CLI, Cursor; Python 3.10+ and pip/uv per README

**Setup time:** 15-35 minutes

### Key facts (verified)

- GitHub: 1216 stars · 249 forks · pushed 2026-05-14.
- License: MIT · owner avatar + repo URL verified via GitHub API.
- README-backed entrypoint: `ag-refresh / ag-ask`.

## Main

- Run `ag-refresh` first: it builds the knowledge base so later Q&A is grounded in module docs rather than random file picks.

- Use `ag-ask` as your default onboarding interface: ask architecture questions and require file-path evidence in answers.

- Keep `.antigravity/` under version control when useful so context stays consistent across IDEs and teammates.

- Enable MCP only when you need tool-style integration; the CLI workflow already works across hosts per README.

### README (excerpt)

<div align="center">

<img src="docs/assets/logo.png" alt="Antigravity Workspace" width="200"/>

**Antigravity — Workspace Template for Claude Code, Codex CLI & Cursor**

### Multi-agent knowledge engine + optional MCP server that turns any repo into a queryable AI assistant. Works with Claude Code, Codex CLI, Cursor, Windsurf, and Gemini CLI.

`ag-refresh` builds a knowledge base. `ag-ask` answers questions. Any LLM, any IDE.

Language: **English** | [中文](README_CN.md) | [Español](README_ES.md)

[![License](https://img.shields.io/badge/License-MIT-green?style=for-the-badge)](LICENSE)
[![Python](https://img.shields.io/badge/Python-3.10+-3776AB?style=for-the-badge&logo=python&logoColor=white)](https://python.org/)
[![CI](https://img.shields.io/github/actions/workflow/status/study8677/antigravity-workspace-template/test.yml?style=for-the-badge&label=CI)](https://github.com/study8677/antigravity-workspace-template/actions)
[![DeepWiki](https://img.shields.io/badge/DeepWiki-Docs-blue?style=for-the-badge&logo=gitbook&logoColor=white)](https://deepwiki.com/study8677/antigravity-workspace-template)
[![NLPM](https://img.shields.io/badge/NLPM-audited-7C3AED?style=for-the-badge)](https://github.com/xiaolai/nlpm-for-claude)

<br/>

<img src="https://img.shields.io/badge/Cursor-✓-000000?style=flat-square" alt="Cursor"/>
<img src="https://img.shields.io/badge/Claude_Code-✓-D97757?style=flat-square" alt="Claude Code"/>
<img src="https://img.shields.io/badge/Windsurf-✓-06B6D4?style=flat-square" alt="Windsurf"/>
<img src="https://img.shields.io/badge/Gemini_CLI-✓-4285F4?style=flat-square" alt="Gemini CLI"/>
<img src="https://img.shields.io/badge/VS_Code_+_Copilot-✓-007ACC?style=flat-square" alt="VS Code"/>
<img src="https://img.shields.io/badge/Codex-✓-412991?style=flat-square" alt="Codex"/>
<img src="https://img.shields.io/badge/Cline-✓-FF6B6B?style=flat-square" alt="Cline"/>
<img src="https://img.shields.io/badge/Aider-✓-8B5CF6?style=flat-square" alt="Aider"/>

</div>

<br/>

<div align="center">
<img src="docs/assets/before_after.png" alt="Before vs After Antigravity" width="800"/>
</div>

<br/>

## Why Antigravity?

> An AI Agent's capability ceiling = **the quality of context it can read.**


### Source-backed notes

- README documents the core commands `ag-refresh` and `ag-ask`, and shows manual install via `pip install git+...#subdirectory=cli/engine`.
- README claims a head-to-head eval on 36 questions and includes quantitative results and dates (see README section headings).
- Repo is MIT licensed and supports multiple hosts (Claude Code, Codex, Cursor) per README badges/table.

### FAQ

- **Do I need MCP enabled?**: No—README shows CLI-only flows; MCP is optional when you want tool integration.
- **What are the core commands?**: `ag-refresh` builds/refreshes knowledge; `ag-ask` answers questions over the current workspace (README).
- **How long does a first setup take?**: Expect ~15–35 minutes depending on model/provider setup and repo size; tune timeouts in `.env`.

## Source & Thanks

> Created by [study8677](https://github.com/study8677). Licensed under MIT.
>
> [study8677/antigravity-workspace-template](https://github.com/study8677/antigravity-workspace-template) — ⭐ 1216

Thanks to the upstream maintainers and contributors for publishing this work under an open license.

---

<!-- ZH -->

## Quick Use

```bash
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=cli"
pip install "git+https://github.com/study8677/antigravity-workspace-template.git#subdirectory=engine"
cat > .env <<EOF
OPENAI_BASE_URL=https://your-endpoint/v1
OPENAI_API_KEY=your-key
OPENAI_MODEL=your-model
AG_ASK_TIMEOUT_SECONDS=120
EOF
ag-refresh --workspace .
ag-ask "How does auth work in this project?"
```

## Intro

Antigravity 用 `ag-refresh` 构建知识库、用 `ag-ask` 做仓库问答，并可选启用 MCP server；适合把任意代码库变成可查询助手；已验证 1216★，更新于 2026-05-14。

**Best for:** 需要快速上手大型仓库，并希望回答带文件路径证据与路由能力的团队

**Works with:** Claude Code、Codex CLI、Cursor；README 要求 Python 3.10+ 与 pip/uv

**Setup time:** 15-35 minutes

### Key facts (verified)

- GitHub：1216 stars · 249 forks；最近更新 2026-05-14。
- 许可证：MIT；作者头像与仓库链接均已通过 GitHub API 复核。
- README 中可对照的入口：`ag-refresh / ag-ask`。

## Main

- 先跑 `ag-refresh`：先把知识库建起来，后续问答才会基于模块文档而不是随机翻文件。

- 把 `ag-ask` 当默认上手入口：问架构问题时要求回答带文件路径证据，降低幻觉率。

- 需要时把 `.antigravity/` 纳入版本控制，让 IDE 与团队看到一致上下文。

- 只有在需要“工具式集成”时再启用 MCP；按 README CLI 流程已可跨宿主使用。

### README (excerpt)

<div align="center">

<img src="docs/assets/logo.png" alt="Antigravity Workspace" width="200"/>

**Antigravity — Workspace Template for Claude Code, Codex CLI & Cursor**

### Multi-agent knowledge engine + optional MCP server that turns any repo into a queryable AI assistant. Works with Claude Code, Codex CLI, Cursor, Windsurf, and Gemini CLI.

`ag-refresh` builds a knowledge base. `ag-ask` answers questions. Any LLM, any IDE.

Language: **English** | [中文](README_CN.md) | [Español](README_ES.md)

[![License](https://img.shields.io/badge/License-MIT-green?style=for-the-badge)](LICENSE)
[![Python](https://img.shields.io/badge/Python-3.10+-3776AB?style=for-the-badge&logo=python&logoColor=white)](https://python.org/)
[![CI](https://img.shields.io/github/actions/workflow/status/study8677/antigravity-workspace-template/test.yml?style=for-the-badge&label=CI)](https://github.com/study8677/antigravity-workspace-template/actions)
[![DeepWiki](https://img.shields.io/badge/DeepWiki-Docs-blue?style=for-the-badge&logo=gitbook&logoColor=white)](https://deepwiki.com/study8677/antigravity-workspace-template)
[![NLPM](https://img.shields.io/badge/NLPM-audited-7C3AED?style=for-the-badge)](https://github.com/xiaolai/nlpm-for-claude)

<br/>

<img src="https://img.shields.io/badge/Cursor-✓-000000?style=flat-square" alt="Cursor"/>
<img src="https://img.shields.io/badge/Claude_Code-✓-D97757?style=flat-square" alt="Claude Code"/>
<img src="https://img.shields.io/badge/Windsurf-✓-06B6D4?style=flat-square" alt="Windsurf"/>
<img src="https://img.shields.io/badge/Gemini_CLI-✓-4285F4?style=flat-square" alt="Gemini CLI"/>
<img src="https://img.shields.io/badge/VS_Code_+_Copilot-✓-007ACC?style=flat-square" alt="VS Code"/>
<img src="https://img.shields.io/badge/Codex-✓-412991?style=flat-square" alt="Codex"/>
<img src="https://img.shields.io/badge/Cline-✓-FF6B6B?style=flat-square" alt="Cline"/>
<img src="https://img.shields.io/badge/Aider-✓-8B5CF6?style=flat-square" alt="Aider"/>

</div>

<br/>

<div align="center">
<img src="docs/assets/before_after.png" alt="Before vs After Antigravity" width="800"/>
</div>

<br/>

## Why Antigravity?

> An AI Agent's capability ceiling = **the quality of context it can read.**


### Source-backed notes

- README 记录核心命令 `ag-refresh` 与 `ag-ask`，并给出通过 `pip install git+...#subdirectory=cli/engine` 的手动安装方式。
- README 包含 36 个问题的对比评测章节，并给出量化结果与日期（见对应标题）。
- 仓库为 MIT 许可证，README 标记支持 Claude Code、Codex、Cursor 等多宿主。

### FAQ

- **必须开 MCP 吗？**：不必须：README 给出纯 CLI 流程；只有需要工具式集成时再开 MCP。
- **核心命令是什么？**：`ag-refresh` 构建/刷新知识库，`ag-ask` 基于当前工作区做问答（见 README）。
- **首次配置大概要多久？**：通常 15–35 分钟，取决于 Provider/模型配置与仓库规模；可在 `.env` 调超时。

## Source & Thanks

> Created by [study8677](https://github.com/study8677). Licensed under MIT.
>
> [study8677/antigravity-workspace-template](https://github.com/study8677/antigravity-workspace-template) — ⭐ 1216


---
Source: https://tokrepo.com/en/workflows/antigravity-repo-q-a-engine-optional-mcp
Author: MCP Hub