OpenDemon

Claude省钱版 — 效果基本一致的情况下可以节约Token消耗约90%,超级省Token,支持OpenAI及国产大模型(DeepSeek/GLM/通义千问)的AI编程助手。本项目基于ClaudeCode源码实现。

14
7
100% credibility
Found Apr 03, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

Claude-ACE is a token-efficient AI coding assistant that supports multiple AI models to analyze, edit, test, and self-heal codebases.

How It Works

1
🔍 Discover Claude-ACE

You hear about this smart coding sidekick that helps fix and improve your code while saving tons of money on AI help.

2
📥 Get it ready

Download it to your project folder and start it up with a simple launch.

3
🔑 Link your AI helper

Follow the easy guide to connect a smart AI service like GLM or others, and it remembers your choices for next time.

4
Trust your project

Quickly confirm it's okay to work on your code files, keeping everything safe.

5
💬 Ask for code help

Type what you need, like 'fix this bug' or 'make it faster', and watch it go to work.

6
🛠️ Magic fixes happen

It smartly peeks at your files, runs checks, heals issues, and suggests better ways – all super efficient.

🎉 Perfect code achieved

Your project runs smoothly with bugs gone, better design, and you saved 90% on AI costs – ready for more wins!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is claude-ace?

Claude-ace is a Node.js CLI AI coding assistant that delivers Claude-level code understanding while slashing token costs by 90% on large files. It feeds models smart AST skeletons instead of dumping full code, expanding only needed functions on demand, and works with OpenAI or Chinese providers like DeepSeek, GLM, and Qwen via simple env vars. Run ace globally for a REPL that reads/writes files, runs shell commands, and verifies changes with auto-generated tests.

Why is it gaining traction?

Token benchmarks prove 84-90% savings on real tasks like architecture analysis, beating full-context tools without quality loss. Standouts include runtime model switching (/model deepseek-chat), cross-project memory for debugging lessons, call graph impact analysis before refactors, and a watchdog daemon that runs npm test to auto-fix breaks. Safety blocks risky bash/file ops, making claude ace github integration feel production-ready.

Who should use this?

Node/TS devs refactoring monorepos tired of token bills on gpt-4o-mini, backend teams needing cheap DeepSeek-powered code reviews, or solo fullstackers prototyping features with self-healing tests. Ideal for claude github copilot alternatives on domestic models like GLM-4-flash (free tier).

Verdict

Try it for claude ace token savings on JS projects—global ace command and /setup wizard get you coding fast. With 14 stars and 1.0% credibility, it's pre-1.0 but battle-tested via benchmarks and v0.8.5 security patches; monitor for stability in prod.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.