garyqlin

garyqlin / glink-engine

Public

Zero-dependency multi-agent workflow orchestration engine. YAML pipelines, shared event bus, auto-recovery, real-time dashboard.

github.comgaryqlinglink agent-orchestration multi-agent pipeline python workflow-automation
20
0
80% credibility
Found May 29, 2026 at 21 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Glink is a multi-agent workflow orchestration engine that lets you describe a project goal in plain language, then automatically coordinates multiple AI assistants to build it step by step. It handles task sequencing, automatic retries when things fail, real-time progress tracking via a dashboard, and can resume interrupted work from where it left off. The system supports different types of steps (regular tasks, code reviews, compact summaries, and shell commands with security restrictions), and can send progress updates to messaging channels like Feishu or the console.

How It Works

1
💡 You have an idea for a project

You think of something you want to build - maybe a game, a website, or a tool - and you want AI helpers to work on it together.

2
📋 You describe what you want

You write a simple one-sentence description of your project goal, and the system automatically plans out all the steps needed to build it.

3
🤖 AI helpers spring into action

Different specialized AI assistants each take on their assigned part - one might design the look, another writes the code, another checks the work - all working in sequence or together.

4
📊 You watch progress in real time

A live dashboard shows you exactly what's happening right now - which step is running, what's been completed, and if anything needs attention.

5
Something goes wrong or everything works?
🔄
Automatic recovery kicks in

The system notices the problem, tries again, or switches to a backup helper without bothering you

✅
Everything keeps flowing

Each step completes successfully and the next one starts automatically

🎉 Your project is complete

All steps finish and you get your finished project - whether it's a working game, a functional app, or whatever you imagined. You can restart anytime to pick up where you left off.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 21 to 20 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 glink-engine?

Glink-engine is a Python-based workflow orchestration engine for coordinating multiple AI agents. You define pipelines in YAML, and the engine handles sequencing, retries, dependency checking, and event tracking across your agent fleet. It ships with a real-time dashboard API that streams workflow progress via Server-Sent Events, plus a shared event bus that gives every agent visibility into what others have accomplished.

Why is it gaining traction?

The checkpoint-and-resume mechanism is the real hook. If a workflow crashes mid-execution, glink-engine picks up exactly where it left off without re-running completed steps. Combined with the fallback agent routing (if your coding agent goes offline, it automatically routes to a backup), this handles production chaos without manual intervention. The YAML-first approach means non-technical stakeholders can read and audit workflows without touching code.

Who should use this?

Teams running multiple AI agents in sequence or parallel will get the most value. If you are building products that chain agents together (code generation into review into testing), glink-engine replaces ad-hoc scripts with something observable and restartable. The dashboard is useful for team leads who need visibility into long-running agent pipelines without digging through logs.

Verdict

At 20 stars and v0.3.3, glink-engine shows promise but carries the credibility score of 0.800000011920929%, reflecting an early-stage project with limited community validation. The feature set is solid for teams already running a local agent fleet on HTTP ports, but the hardcoded agent port assumptions mean it works best within a specific internal setup. Evaluate it for pilot projects before committing to production workflows.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.