I Built an Open Engine That Connects Claude, ChatGPT, and Codex Together
TL;DR
The human becomes the hallway: When agents live in separate rooms, you carry the work between them, and that invisible labor is exhausting.
A queue is the missing coordination layer: Put work in a ticket system (Linear, Jira, anything agents can read/write) and the ticket becomes the place agents talk.
Prompt mode asks for answers, work mode gives jobs: A prompt says "write me an email" while a ticket says "here's the transcript, constraints, and definition of done, flag what needs my judgment."
Agents from different providers can delegate to each other: Maya's Codex can route work to Leo's Claude by writing a self-contained Linear issue with full context.
The receipt matters as much as the work: Agents must leave proof of what they did and where they stopped so the next person or agent can pick it up without reading chat transcripts.
This works for households and teams: School pickup changes and sales pipelines have the same handoff shape, and Open Engine handles both.
The Breakdown
Nate B Jones built Open Engine, a queue-based system that connects Claude, ChatGPT, Codex, and other AI agents through a shared task system so work can flow between them without humans acting as the copy-paste glue. The core insight is that the real bottleneck isn't model capability, it's handoffs between agents, and a simple Linear ticket queue with the right protocol solves this.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
The Cheapest Model That Passes
OpenRouter lists 400 models behind one API. The fix for choosing isn't a better leaderboard, it's a four-step protocol that ends in a real eval.
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.
Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.