Analysis updated 2026-05-18
Give an AI agent a local project folder and watch it reason through and complete a task.
Review a diff of every file change the agent makes before accepting it.
Manage multiple chat sessions across different project directories.
Set up background automation with Claw to handle messages from Feishu or scheduled tasks.
| xingyu-zhong/deepseek-gui | alpha-vector-core/polymarket-arbitrage-trading-bot | alpha-vector-core/polymarket-trading-bot | |
|---|---|---|---|
| Stars | 79 | 79 | 79 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 4/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires your own DeepSeek API key, Linux has no pre-built installer and must be built from source.
This is a desktop application that wraps the DeepSeek AI agent tool in a graphical interface, making it easier to use for people who do not want to work in a terminal. The README is written in Chinese, with an English version also available in the repository. It is built on top of an existing terminal-based tool called DeepSeek TUI and packages that experience into a windowed application for macOS and Windows, with Linux support via building from source. The main purpose is to give developers and frequent AI users a persistent workspace where they can assign tasks to an AI agent, watch it reason through problems, see which files it touches, and approve or reject actions it wants to take. You pick a local project folder, describe what you want done, and the application shows you the agent's thought process, tool calls, and any file changes in the same window. A side panel displays a diff of every modification so you can review changes before accepting them. The application manages multiple chat sessions, each associated with a specific working directory. You can switch between projects, review past conversations, and interrupt or retry a generation at any time. Permission levels can be configured to limit what the agent is allowed to do, ranging from read-only access up to full file system control. Certain tool actions can be set to require manual approval before execution. A separate mode called Claw handles background automation. It can connect to Feishu (also known as Lark, a Chinese workplace messaging platform), receive messages there, and have an AI agent respond or take action. It also supports scheduled tasks and local webhook endpoints. The application stores all settings, session history, and runtime configuration locally on your machine. It uses your own DeepSeek API key for model calls. Pre-built installers are available for macOS (.dmg) and Windows (.exe) on the GitHub releases page.
A desktop app that wraps the DeepSeek AI agent in a graphical window, letting you assign coding tasks, watch it reason, and review file changes before approving them.
Mainly TypeScript. The stack also includes TypeScript, Electron.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.