loopyard

Multiple agents per project

Run setup, feature work, and a debugger on the same codebase at the same time.

The same Loopyard session on a phone: Multiple agents per project

Three agents on one codebase. One is mid-fix, one is running the suite, one is waiting for review work. Each has its own live status in the rail.

A codebase rarely has one job in flight. Something is being written, a bug is being chased, a migration needs prepping, the test suite needs hardening. One agent turns all of that into a queue.

Loopyard lets you spin up several agents against the same project at the same time. One is writing a new endpoint while another is reproducing the flaky test that breaks in CI; a third is running the formatter across the diff; a fourth is reading the migration to draft a rollback. They share the same workspace, the same database, the same dev server, the same view of the world.

Each agent has its own chat, terminal, and scratch space. Watch them all, talk to the one that matters right now, ignore the rest. When two agents need the same file, that call is yours to make. Loopyard does not pretend to have solved a coordination problem it has not solved.

Run it on your machine

git clone github.com/loopyard/loopyard.git
cd loopyard
mix loopyard.setup

Free and open source on GitHub. Follow @loopyardai.