MoQ Monthly, Issue #1

April 2026 (barely)

I said monthly. I meant roughly monthly. NAB happened. Between prepping demos, traveling to Vegas and processing all of the good conversations had there, April really got away from me. To make up for it, I have a lot to cover in this issue!

NAB 2026

The headline story from the past couple months is NAB Show 2026 (April 18-22, Las Vegas), and when it comes to MoQ, the headline story from NAB is interoperability. Not just demos, not just announcements, but broad interoperability. A working, multi-vendor MoQ ecosystem showed up on the show floor, and for a protocol that has sometimes been described as promising but fragmented, that was really exciting to see and be a part of.

For me, the clearest illustration of this came from a last-minute addition to a demo that Qualabs helped build. Nicolás Levy shared the story on LinkedIn: Qualabs, Ateme, and EZDRM had already put together a demo combining C2PA provenance signing, DRM, and MoQ delivery. When Nicolás posted about it before the show, I got excited. After hearing "DRM" called out so many times as something people hadn't seen done with MoQ yet, here it was! Plus, C2PA! Content provenance is increasingly important and seeing all of this together working over MoQ felt like a significant turning point. I also noticed that the diagram had a "relay" but no CDN to provide capacity for fan out at scale. So, I DM'd him - "Love this! Can we be the relay?" A few hours later, Qualabs, EZDRM, and Ateme had all said yes, and we had the demo running end-to-end across all these organizations and through Cloudflare's global relay network. As Nicolás noted, integration meant pointing the player to a different MOQT URL and configuring Cloudflare's relay network to appropriately pull from their origin. That was it. No hacks, no workarounds, no bilateral negotiation. That's what a working standard looks like in practice.

I've been lucky to work with the Qualabs team across multiple MoQ Summer Projects now, and they bring exactly this kind of collaborative, relationship-first energy to everything they do and the technical results they achieve are hard to argue with. The full demo writeup is worth reading.

Elsewhere on the floor, MoQ was genuinely hard to miss. A few things worth singling out:

Oracle Video @ Edge ran what has been described as the first confirmed public multi-vendor MoQ interop demo at any show: Ateme encoding into Oracle's relay fabric, Broadpeak handling packaging, Cloudflare and Broadpeak providing CDN delivery, and Bitmovin's Player Web X at the endpoint. Oracle's blog post covers the architecture in detail. The Bitmovin writeup is also good, particularly on why interoperability is the whole point rather than any single vendor's implementation.

Wowza demoed a stack running OBS into Wowza origin infrastructure, CMAF packaged over MoQT in CMSF format, relayed through Cloudflare, and played back in an experimental Shaka Player build. Their blog post has details. Shaka Player support for MoQ is also noteworthy - this is a widely deployed player used in many production deployments. For existing Shaka users, being able to start using MoQ without switching to a completely different player has some real appeal.

Norsk (by id3as) demonstrated native MoQ support in a joint demo at the Cloudflare booth, with MoQ integrated as the default protocol for preview paths and latency measurement built into the workflow. Their CTO Steve Strong has been really thoughtful about what MoQ enables.

The OpenMOQ Software Consortium hosted a Streaming Summit panel session to a room so full it had Dan Rayburn scrambling to get more seats. Will Law (Akamai) moderated, with Arvind Suryakumar (Oracle), Chris Allen (Red5), Cullen Jennings (Cisco), Gwendal Simon (Synamedia), Sean McCarthy (YouTube), and Tomas Kvasnicka (CDN77) as speakers. I don't think the recordings from the Streaming Summit are available just yet, but the session page has speaker details.

We also did some interesting integrations at the Cloudflare booth with Jeket (demoing their interactive broadcast suite remotely) and Hydrolix (streaming moqt-event qlog structured logs directly from our relay into Hydrolix and using their AI tooling to automatically ingest the data and build Grafana dashboards - a good sign for MoQ observability). Beyond that, there was a whole lot more MoQ-related stuff I didn't have time to see in person: Synamedia, Ant Media, Broadpeak, Fraunhofer FOKUS, Red5, nanocosmos, and more.

For the most complete picture of what was shown across the floor, Nicolas Weil's NAB Field Guide maps it all out by protocol stack layer - the place to start if you want the full story.

AVEQ also published a short video roundup featuring Werner Robitza talking to Luke Curley, Steve Strong, Dom Robinson (Norsk), Dr. Louay Bassbouss (Fraunhofer FOKUS), and me about where MoQ stands - a nice snapshot of the conversation at NAB.

A couple things from February I didn't cover in Issue #0

At ACM Mile High Video in Denver (Feb 2-5), Broadpeak presented work on "Streaming large-scale live events: transitioning to MoQ" (day 3, Video Streaming 6 track). The full technical program has the details. The technical substance that generated the most hallway discussion was an observation about where the real hardware opportunity in MoQ lies: QUIC stacks today carry more CPU overhead than TCP stacks that have been optimized for decades (especially when it comes to TLS offload), but, interestingly, MoQ's pub/sub model opens up acceleration opportunities that HTTP simply doesn't have. In HTTP, a CDN cache has no advance knowledge of who will request content next. In MoQ, SUBSCRIBE gives the relay explicit, advance knowledge of each subscriber and the objects flow sequentially and predictably as continuously delivered tracks rather than discretely requested segments. Fan-out is knowable. That structural property is different from anything the existing HTTP caching model supports, and the implications for hardware-accelerated relay architectures aren't fully worked out yet.

Marc Fiuczynski (formerly a Principal Architect at Akamai, now a visiting scholar at Rutgers CS / WINLAB) attended that session and we talked through the implications over lunch at the MSF workshop the next day. He subsequently wrote it up: "MoQ Relay Has a Hardware Problem. Is Anyone Connecting the Dots?" Marc and Richard Martin are now exploring SmartNIC primitives at Rutgers that could address this specifically. A good example of how conference presentations and hallway conversations can end up turning into real research directions.

Also from February: Vindral and SVT published a press release confirming that Vindral's MoQ-based platform was part of SVT's remote production setup for the Milano Cortina Winter Olympics, as part of SVT's NEO (Next-Gen Online Production) initiative. MoQ for contribution at the Winter Olympics is worth noting, even if viewer-facing delivery wasn't in scope for that particular project. Streaming Media's piece on their 2025 World Rally Championship trial covers more of Vindral's MoQ backstory.

Interestingly, using MoQ as a means of moving high quality streams around on the production side of the house was also a theme of several conversations I had at NAB this year. We've often focused on contribution and distribution, but it's seeming more and more like MoQ might prove especially useful for some middle-mile use cases as well.

Nicolas Weil also published a two-part architectural reassessment of streaming workflows in early March with a thorough and independent treatment of where MoQ fits: Part 1 covers the HTTP streaming status quo; Part 2 is where MoQ shows up.

Spec Status

Draft-14 has been fairly widely implemented and we've been working on draft-16 interop as well. The moq-interop-runner and public test streams are good tools for validating interoperability. We're looking ahead towards a forthcoming draft-18 as the next interop target. For a current summary of all active drafts and their status, the MoQ LLM Wiki (more on that below) has a useful table.

Browser Support

Safari 26.4 shipped WebTransport without requiring developer mode flags, making WebTransport Baseline across all major browsers including Safari on iOS. This writeup from WebRTC.ventures explains what that means practically for MoQ. The iOS angle matters a lot given that every iPhone browser was previously affected by Safari's lack of WebTransport support. There are still some caveats - WebCodecs maturity on Safari and some transferable streams details - so it's not quite "MoQ in every browser with no fallbacks" yet. But meaningfully better.

Community

OpenMOQ added Vindral as a member in late February and has moved into what they describe as the core infrastructure development stage, with legal structure and governance finalized. The C++ relay work is still the main deliverable from what I can tell, and it looks like the earliest version of that work is now up on GitHub as moqx.

On the Cloudflare side, we posted on LinkedIn what we've been up to. Two community tools worth highlighting specifically: the moq-interop-runner now automatically tests against moq.dev, imquic, Meta's moxygen, Cisco's libquicr, Google's quiche, MOQtail, Alibaba's xquic, Eyevinn's moqlivemock, OpenMOQ's moqx, and more, with public results. And MoQ Bench is a new tool in the Cloudflare dashboard (Media > Realtime > MoQ Bench) that lets any MoQ implementer spin up a whole bunch of real subscribers distributed across Cloudflare's network to stress-test their setup. Public test streams are also now available running through our relay network, and the Bitmovin Player Web X test player can connect to them.

Giovanni Marzot (Marz Research) has resumed work on aiomoqt, his Python asyncio MoQT implementation, now targeting draft-14 and draft-16. A Python MoQT implementation is a natural fit for testing and flexible experimentation. I'm curious to see if this makes MoQT more accessible to a broader audience of experimenters and developers.

Torbjörn Einarsson (Eyevinn Technology) has shipped a significant update to moqlivemock and warp-player: draft-14 and draft-16 support with auto-negotiation, DRM (Widevine, PlayReady, FairPlay, ClearKey) following CMSF PR 18, and confirmed working WebTransport on Safari 26.4 including iOS. He's also contributed a Docker image with an mlmtest component to the moq-interop-runner as a PR.

Torbjörn is also running an experiment using Andrej Karpathy's LLM Wiki concept to build a living MoQ ecosystem reference, updated daily from the mailing list, Slack, GitHub, and Datatracker. It's early - the quality of automatically-generated content will vary - but it seems like it really could be pretty useful. Worth bookmarking and watching to see how it develops.

Coming Up

Streaming Tech Sweden is May 21 in Stockholm. Vindral's CTO Per Mafrost is presenting on their end-to-end MoQ implementation experience, and there's a session on SVT's Olympics production setup. Very much worth attending if you're in the Nordics. (I spoke about MoQ at Streaming Tech Sweden last year and was very impressed by the event.)

The next IETF MoQ WG hybrid interim is June 9-12 in London, hosted at Cloudflare's offices. The initial announcement is on the list - I'll send more logistical details to the IETF MoQ WG mailing list soon.

Resources

• IETF MoQ WG: datatracker.ietf.org/group/moq
• Mailing list: mailarchive.ietf.org/arch/browse/moq
• MoQ.dev community Discord: discord.gg/FCYF3p99mr

If you're working on something MoQ-related - a draft, an implementation, a deployment, a talk, a blog post - I'd love to hear about it and help share your news with the broader community.

Subscribe: buttondown.com/moqmonthly

See you in May!

-Mike