Adamlixi

Adamlixi / codenano

Public

Lightweight AI coding agent SDK extracted from Claude Code. 5,500 lines vs 510,000+. Same power, 97% less complexity. Build custom AI coding agents in 1 minutes

agent claude-agent-sdk claude-code
51
14
100% credibility
Found Apr 03, 2026 at 51 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

codenano is a lightweight open-source SDK for quickly building customizable AI coding agents with built-in tools for file operations, shell commands, memory, and session persistence.

How It Works

1
🔍 Discover codenano

You hear about a simple tool that lets anyone build a smart coding helper, like having an AI sidekick for your projects.

2
📦 Add it easily

You bring it into your project with one quick step, no hassle or complicated setup.

3
🧠 Create your helper

You connect a thinking AI and pick some basic abilities like reading or changing files, and your personal coding assistant comes alive.

4
đź’¬ Give it a job

You ask it something simple, like 'summarize this file' or 'fix this code', and watch it work step by step.

5
🔄 Chat back and forth

It remembers what you talked about before, so you can pick up right where you left off, even days later.

🚀 Code smarter every day

Your reliable AI buddy handles boring tasks, remembers your style, and helps you build faster without the headaches.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 51 to 51 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 codenano?

codenano is a TypeScript SDK for building lightweight coding agents that handle file reads, edits, bash execution, git state detection, and web searches out of the box. npm install it, spin up an agent with a Claude model and core tools in three lines, and get production features like session persistence, cross-session memory, cost tracking in USD, and streaming tool execution. At 5,500 lines versus bloated alternatives' 510,000+, it delivers Claude Code's agent loop—auto-compaction, tool budgeting, lifecycle hooks—without IDE ties or UI overhead.

Why is it gaining traction?

It slashes complexity for AI coding agents: zero-config persistence to JSONL files, MCP protocol for GitHub tools, sub-agent spawning, and 17 ready tools like LSP and notebook edits. Developers dig the battle-tested reliability—retry logic, 413 recovery, max-output escalation—over generalist SDKs like Vercel AI or LangChain's kitchen sink. Quick wins like git prompt injection and memory extraction make it a lightweight coding agent that runs in your terminal, hooking indie hackers building custom bots.

Who should use this?

Backend engineers embedding agents in CLI tools or scripts, startups prototyping AI code reviewers without framework lock-in, or ops teams needing git-aware automation. Ideal for TypeScript devs tired of heavyweight RAG setups, wanting a lightweight coding LLM wrapper with hooks to block risky bash commands—perfect if you're evaluating github lightweight alternatives for agentic workflows.

Verdict

Grab it if you need a focused, lightweight coding agent SDK now—docs shine, 374 tests pass, MIT license. But 51 stars and 1.0% credibility score scream early days; production use demands your own hardening. Promising for lightweight coding software stacks, worth a spin over verbose options. (198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.