he-yufeng

he-yufeng / AnyCoder

Public

AI coding agent in your terminal. Works with any LLM - DeepSeek, Qwen, GPT-5, Claude, Gemini, Kimi, Ollama. 100+ models via litellm. ~1300 lines of Python.

11
1
100% credibility
Found Apr 08, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

AnyCoder is a terminal-based AI assistant that helps with coding tasks like reading files, making edits, running commands, and searching codebases using any supported language model.

How It Works

1
🔍 Discover AnyCoder

You hear about a handy coding helper that lives right in your terminal and works with any smart AI you like.

2
📦 Get it set up

You add this simple tool to your computer so it's ready to use anytime.

3
🧠 Pick your AI helper

You link it to your favorite smart AI service, like a quick one for everyday tasks or a powerful one for big jobs.

4
💬 Chat about your code

You type a plain request like 'read main.py and fix the broken import,' and it springs into action.

5
👀 Watch the magic

It peeks at your files, makes smart edits, shows you the exact changes with a clear before-and-after, and even runs safe checks.

6
💾 Save your progress

You can pause anytime, save the conversation, and pick right back up later with a quick restart.

Code perfected

Your project is fixed, running smoothly, and you've got a reliable sidekick for all future coding adventures.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 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 AnyCoder?

AnyCoder is a terminal-based AI coding agent that hooks into any LLM—DeepSeek, Qwen, GPT-5, Claude, Gemini, or 100+ more via litellm—to read files, edit code with diffs, run safe shell commands, and search codebases. Built in ~1,300 lines of Python, it delivers Claude Code-like workflow without API lock-in: stream responses, persist sessions with `/save`, and resume via `--resume`. Pip install and run `anycoder -m deepseek` to fix imports or refactor instantly.

Why is it gaining traction?

In coding agents comparison and coding agents ranking, AnyCoder shines with universal LLM support, dodging vendor pricing traps—use cheap DeepSeek or local Ollama for zero-cost runs. User-facing perks like parallel tools, context compression for long chats, and diff previews make coding agent workflow smooth, while bash safety blocks rm -rf disasters. Devs grab it for the lean, hackable github coding agent that's MIT-licensed and beats Aider/Cline on model flexibility.

Who should use this?

Python backend devs tweaking APIs or debugging CLI tools without IDE swaps. Open source maintainers prototyping PRs across repos, or Chinese teams using Qwen for bilingual coding agent skills. Solo hackers benchmarking coding agent open source options on Raspberry Pi setups or personal github coding agent jobs.

Verdict

Worth pip-installing for flexible anycoder AI in your terminal—11 stars and 1.0% credibility score signal early-stage freshness, but solid docs, tests, and .env setup make it reliable for prototypes. Pair with DeepSeek for cheap coding agent github experiments; hold for enterprise until stars climb.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.