Skip to content
@treqs

TReqs

TReqs (sounds like T-Rex) helps you plan, ship, and govern AI work on your own infrastructure.

Capture what ran. Store the record. Control what runs next.

TReqs gives you a complete record of how your models are actually built — and what gets to run next.

roarIf it ran, roar saw it.

A CLI that captures lineage at runtime — data, code, environment, artifacts. No code changes. No loggers. No frameworks. Maximal visibility.

GLaaSThe registry of roar activity.

A content-addressable registry of every artifact and job. Resolve any artifact's hash back to the code, data, and environment that made it. -> GLaaS Docs <-

TReqsApprove before compute.

Training requests as pull requests. As much or as little peer review as your team needs for important compute jobs. Nothing runs until someone — or a policy — says go. -> TReqs Docs <-

TReqs Tools

roar Use roar to record what was read, what was written, the commit, and the environment.
treqs-cli Drive TReqs from a terminal and with the agent-friendly CLI.

Pinned Loading

  1. roar roar Public

    Run Observation & Artifact Registration

    Python 5

  2. treqs-cli treqs-cli Public

    CLI for the TReqs platform

    Python 1

Repositories

Showing 10 of 11 repositories

Top languages

Loading…

Most used topics

Loading…