# Refact — Local-First AI Coding Assistant

> Refact is an open-source, local-first AI coding assistant: install the IDE plugin, run local refact-lsp, and connect a model provider.

## Install

Save as a script file and run:

## Quick Use

1. Install an IDE plugin (see docs):
   - VS Code: https://docs.refact.ai/installation/vs-code/
   - JetBrains: https://docs.refact.ai/installation/jetbrains/
2. Open a workspace and launch Refact; it starts the local `refact-lsp` engine.
3. In Provider Setup, connect a provider/runtime and pick default models for chat + agent work.

## Intro

Refact is an open-source, local-first AI coding assistant: install the IDE plugin, run local refact-lsp, and connect a model provider.

- **Best for:** developers who want a local agent engine embedded in the editor, with BYO providers and repeatable workflows
- **Works with:** VS Code or JetBrains; local `refact-lsp` engine; multiple LLM providers/runtimes (see README)
- **Setup time:** 10–30 minutes

## Practical Notes

- Quant: the repo describes an IDE-embedded flow where the plugin runs a local `refact-lsp` engine per workspace.
- Quant: validate setup by running one agent task end-to-end and measuring time saved over 3 repeated tasks (same prompt, same repo).

## What to standardize before rollout

Refact becomes much more valuable when teams standardize:

1. **Provider policy**: which providers are allowed for which repos (open-source vs private).
2. **Default models**: one for chat, one for agent work, one for embeddings if needed.
3. **Task boundaries**: define which actions require explicit approval (deps updates, migrations, deploy scripts).

## Suggested first workflows

- “Explain module X” + “write a unit test for function Y”.
- “Refactor a file” with a measurable constraint (max 10 lines changed; no behavior change).
- “Fix a failing test” with reproduction steps and a time budget.

Use the same 2–3 workflows across the team so you can compare outcomes consistently.

### FAQ

**Q: Is it only for chat?**
A: No. The README positions it as an agent that can plan, execute, and iterate in engineering workflows.

**Q: Do I have to use one provider?**
A: No. It supports multiple provider families; choose per your policy.

**Q: How do I avoid risky changes?**
A: Define approval-gated actions and start with read/modify-only tasks before automating merges/deploys.

## Source & Thanks

> Source: https://github.com/smallcloudai/refact
> License: BSD-3-Clause
> GitHub stars: 3,541 · forks: 310

---

<!-- ZH -->

## 快速使用

1. 安装 IDE 插件（见文档）：
   - VS Code：https://docs.refact.ai/installation/vs-code/
   - JetBrains：https://docs.refact.ai/installation/jetbrains/
2. 打开项目并启动 Refact；插件会运行本地 `refact-lsp` 引擎。
3. 在 Provider Setup 配置模型提供方/本地运行时，并选择 chat 与 agent 的默认模型。

## 简介

Refact 是开源的本地优先 AI 编程助手：安装 IDE 插件后由本地 `refact-lsp` 引擎驱动，再配置模型提供方，即可在编辑器里跑问答、重构与端到端 agent 工作流等场景。

- **适合谁：** 希望在编辑器内运行本地 agent 引擎、并自选模型提供方的开发者
- **可搭配：** VS Code 或 JetBrains；本地 `refact-lsp` 引擎；多种 LLM 提供方/运行时（见 README）
- **准备时间：** 10–30 分钟

## 实战建议

- 量化信息：README 描述了“插件 + 本地 `refact-lsp` 引擎”的工作方式，按 workspace 启动。
- 量化信息：用同一任务重复 3 次，比较接入前后耗时，作为是否值得全团队推广的依据。

## 推广前先标准化三件事

Refact 在团队里要“好用且不乱用”，关键是先把以下三点标准化：

1. **Provider 策略**：哪些仓库允许用哪些提供方（开源 vs 私有）。
2. **默认模型**：chat / agent / embeddings（如需）各选一个默认，避免每个人乱配。
3. **任务边界**：哪些动作必须显式批准（依赖升级、迁移脚本、发布脚本等）。

## 推荐的首批工作流

- 解释模块 + 为关键函数补单测。
- 约束性重构（例如最多改 10 行、行为不变）。
- 修复失败用例（给复现步骤与时间预算）。

团队先统一 2–3 个固定流程，才能稳定对比效果。

### FAQ

**它只是聊天插件吗？**
答：不是。README 把它定位为能规划、执行并迭代的工程 Agent。

**必须固定一个 provider 吗？**
答：不必须。支持多种 provider；按团队策略选择即可。

**如何避免高风险改动？**
答：把高风险动作做成审批闸门，先从读取/小改动任务开始。

## 来源与感谢

> Source: https://github.com/smallcloudai/refact
> License: BSD-3-Clause
> GitHub stars: 3,541 · forks: 310


---
Source: https://tokrepo.com/en/workflows/refact-local-first-ai-coding-assistant
Author: AI Open Source