uameer

uameer / heartbeat

Public

The KAIROS pattern, open and model-agnostic. Anthropic built it into Claude Code and never told anyone.

10
4
100% credibility
Found Apr 17, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Heartbeat is an open-source background agent that runs locally to monitor coding projects, uses AI to detect opportunities for action, logs structured learnings in owned files, and consolidates memory daily.

How It Works

1
🔍 Discover Heartbeat

You stumble upon Heartbeat, a clever background buddy that keeps an eye on your coding project and jumps in to help when something needs attention.

2
📥 Download it

Grab the project files and place them in a folder on your computer.

3
Choose a thinking partner
☁️
Online AI

Connect to a cloud-based AI for powerful insights.

🏠
Local AI

Use an AI that thinks right on your own machine.

4
🛠️ Get it ready

Follow a couple of easy preparation steps to set everything up.

5
▶️ Start the magic

Point it to your project folder and launch – now it quietly watches every 30 seconds, spotting issues and acting wisely.

6
📖 Review discoveries

Check the notes it keeps on what it learned and any actions it took.

🚀 Smarter project

Your project steadily improves with proactive help, daily insights, and growing wisdom – all stored safely in your files.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 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 heartbeat?

Heartbeat is a Python-based AI agent that runs in the background of your codebase, ticking every 30-120 seconds to scan for issues like TODOs, stale files, git changes, or Django-specific signals such as failed tests and migrations. Inspired by the leaked KAIROS pattern from Anthropic's Claude Code (search claude heartbeat github or kairos ai github), it queries an LLM—Anthropic, OpenAI, Gemini, or Ollama—to decide if action is needed, logs insights to a local JSONL file you own, and consolidates memory daily via autoDream. No cloud lock-in: set your API key and run `./scripts/run.sh .` for autonomous project monitoring.

Why is it gaining traction?

Unlike chat-based tools like OpenClaw or Anthropic's $0.08/hour Managed Agents, heartbeat acts without prompts—proactively spotting patterns via profiles (generic or django) and CLI flags like `--actions` or `--learn`. Developers dig the model-agnostic swap (e.g., `--provider ollama`), append-only logs, and portable learnings, echoing heartbeat ai erfahrungen on GitHub while dodging vendor costs. It's the open heartbeat monitor github for kairos agent github workflows.

Who should use this?

Backend devs maintaining Django apps, solo founders tracking TODOs across repos, or AI tinkerers testing kairos 3.0 github-style autonomy on local Ollama models. Ideal for heartbeat revival in stalled projects, not high-scale prod (yet).

Verdict

Early alpha with 10 stars and 1.0% credibility—docs are solid, CLI intuitive, but lacks tests, notifications, and webhooks; audit the ~450-line core before prod. Try for personal projects if you're into heartbeat ai kosten-free KAIROS, but wait for PRs if stability matters.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.