MoQ Interop Runner
The MoQ working group's automated interop test harness. Runs a nightly client/relay matrix across MoQT implementations and publishes pass/fail results per draft version.
Category
Developer Tools
License
Open Source
Status
Active
MoQT Draft Support
The MoQ Interop Runner, maintained by Mike English (Cloudflare), executes an automated client-against-relay test matrix across MoQT implementations and publishes the results. It is the closest thing the ecosystem has to an objective scoreboard, and it is the source behind the interop reports presented at IETF meetings.
Results are grouped into control-plane conformance cases (setup, announce, publish-namespace-done, subscribe-error, announce-subscribe, subscribe-before-announce) and end-to-end object delivery, run over both raw QUIC and WebTransport.
Why it is worth reading
Compatibility tolerances are opt-in and annotated in the output rather than silently applied, so a pass is a real pass. Behavioural differences that are not failures — such as a relay that holds SUBSCRIBE_OK until upstream serves — are recorded as notes rather than scored as errors.
The run presented at IETF 126 on 2026-07-24 reported 131 passing, 208 failing, and 11 skipped across 350 total runs, all targeting draft-18.
Ongoing work
PR 95 replaces single-version negotiation with a version-pinned per-draft matrix, covering all version overlaps between implementations and adding a partial result tier alongside pass and fail.