rcarmo

rcarmo / piclaw

Public

I'm going to build my own OpenClaw, with blackjack... and bun!

139
7
100% credibility
Found Feb 18, 2026 at 17 stars 8x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

A secure, isolated workspace for running an AI coding assistant, with optional integration for messaging via WhatsApp.

How It Works

1
🔍 Discover PiClaw

You hear about a safe little playground where your smart coding helper AI can work without messing up your computer.

2
🛠️ Set up your playground

You prepare the safe space by building it once, so it's ready for action.

3
🚀 Launch and step in

You start the playground and walk inside, feeling like you're in a protected workshop.

4
🤖 Wake your AI buddy

You chat with your AI helper right there, telling it how to think and what tools to use for coding magic.

5
Pick your chat style
📱
Phone messages

Scan a quick picture code on your phone to link your WhatsApp chats to the AI.

🖥️
Direct talk

Keep chatting right inside the workshop for instant back-and-forth.

6
Ask for coding help

Send messages like 'build me a game' or schedule repeating tasks, and watch your AI create projects.

🎉 Projects come alive

Your AI handles coding tasks securely, saving your files safely, so you get perfect results every time.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 17 to 139 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 piclaw?

PiClaw runs the Pi Coding Agent in a minimal Docker container on Debian Bookworm, bundling Bun runtime, Homebrew, git, vim, and other dev essentials for isolated AI coding sessions. TypeScript-powered piclaw bridges WhatsApp: scan a QR code, message @pi, and get agent responses for code gen, debugging, or tasks—no direct API keys, as Pi handles providers like GitHub Copilot. Users get persistent volumes for workspace, agent state, SQLite chats, and skills like /skill:setup or /skill:schedule via simple make up/enter commands.

Why is it gaining traction?

It delivers WhatsApp-triggered agent loops with typing indicators, queued processing, cron/interval scheduling, and IPC for mid-run updates, all in a multi-arch GHCR image. Lean setup skips bloat—Pi configs providers freely, Bun zips execution—while volumes ensure chats and projects survive restarts. Devs try it for mobile-first AI coding without local installs, echoing agentbox simplicity with nanoclaw's messaging hook.

Who should use this?

AI experimenters needing a WhatsApp coding agent for on-the-go prototypes or remote debugging. Container users deploying Pi for isolated sessions with GitHub Copilot free access. Hobbyists scripting tasks like "npm i github" flows via scheduled prompts.

Verdict

With 10 stars and 1.0% credibility score, it's early-stage—light tests, no heavy validation—but killer docs and MIT license make it forkable. Grab for WhatsApp Pi fun; wait for stars if you need battle-tested reliability.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.