From d532eb3f3b541bcbd10312c8067c3e1b7c079c6a Mon Sep 17 00:00:00 2001 From: Rick Staa Date: Tue, 18 Aug 2026 11:24:05 +0200 Subject: [PATCH 1/2] docs: list the StreamDiffusion app under external examples Its axes (static, persistent, WebSocket, hour) are already covered here, so per the curated-set policy it ships as its own repo and is linked instead. Co-Authored-By: Claude Opus 5 (1M context) --- README.md | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 255f00e..703b69b 100644 --- a/README.md +++ b/README.md @@ -161,11 +161,12 @@ Each entry lists its `runners` with an `app`, `version`, capacity, and `price_in Apps that integrate the live runner and live in their own repos — production deployments and standalone examples alike. This table is links-only: the code, CI, and support stay with the author. -| Project | What it is | Transport | -| ----------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | ------------------- | -| [daydreamlive/scope](https://github.com/daydreamlive/scope/tree/ja/runner) | Real-time AI video with downloadable LoRA models | WebSocket + trickle | -| [livepeer/api-proxy](https://github.com/livepeer/api-proxy) | Attach several API endpoints dynamically — key storage and request stats for operators | HTTP | -| [Gideonjon/vllm-realtime-livepeer-runner](https://github.com/Gideonjon/vllm-realtime-livepeer-runner) | Real-time speech-to-text — trickle audio in, WebSocket transcript out, with live metrics | WebSocket + trickle | +| Project | What it is | Transport | +| ------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | ------------------- | +| [daydreamlive/scope](https://github.com/daydreamlive/scope/tree/ja/runner) | Real-time AI video with downloadable LoRA models | WebSocket + trickle | +| [livepeer/api-proxy](https://github.com/livepeer/api-proxy) | Attach several API endpoints dynamically — key storage and request stats for operators | HTTP | +| [Gideonjon/vllm-realtime-livepeer-runner](https://github.com/Gideonjon/vllm-realtime-livepeer-runner) | Real-time speech-to-text — trickle audio in, WebSocket transcript out, with live metrics | WebSocket + trickle | +| [livepeer/streamdiffusion-livepeer-runner](https://github.com/livepeer/streamdiffusion-livepeer-runner) | Realtime prompt-driven img2img — a third-party StreamDiffusion server run unmodified, reverse-proxied | WebSocket + MJPEG | **Building one?** Start from [**template-livepeer-runner**](https://github.com/livepeer/template-livepeer-runner) — a working app, client, and compose setup you can run in one command, offchain or on-chain. The examples here are not copyable as-is: each one's `compose.yml` pulls the orchestrator from a shared file one directory up. From 97bff07136acfe81f4152b1c040c6ac925a87cde Mon Sep 17 00:00:00 2001 From: Rick Staa Date: Wed, 19 Aug 2026 17:45:58 +0200 Subject: [PATCH 2/2] fix(docs): name the transport, not the payload format MJPEG is what the output stream carries, not how it is carried, and the row left out HTTP even though the output and prompt endpoints are both plain HTTP. Every other row in the table names transports. Co-Authored-By: Claude Opus 5 (1M context) --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 703b69b..eb8d1ac 100644 --- a/README.md +++ b/README.md @@ -166,7 +166,7 @@ Apps that integrate the live runner and live in their own repos — production d | [daydreamlive/scope](https://github.com/daydreamlive/scope/tree/ja/runner) | Real-time AI video with downloadable LoRA models | WebSocket + trickle | | [livepeer/api-proxy](https://github.com/livepeer/api-proxy) | Attach several API endpoints dynamically — key storage and request stats for operators | HTTP | | [Gideonjon/vllm-realtime-livepeer-runner](https://github.com/Gideonjon/vllm-realtime-livepeer-runner) | Real-time speech-to-text — trickle audio in, WebSocket transcript out, with live metrics | WebSocket + trickle | -| [livepeer/streamdiffusion-livepeer-runner](https://github.com/livepeer/streamdiffusion-livepeer-runner) | Realtime prompt-driven img2img — a third-party StreamDiffusion server run unmodified, reverse-proxied | WebSocket + MJPEG | +| [livepeer/streamdiffusion-livepeer-runner](https://github.com/livepeer/streamdiffusion-livepeer-runner) | Realtime prompt-driven img2img — a third-party StreamDiffusion server run unmodified, reverse-proxied | WebSocket + HTTP | **Building one?** Start from [**template-livepeer-runner**](https://github.com/livepeer/template-livepeer-runner) — a working app, client, and compose setup you can run in one command, offchain or on-chain. The examples here are not copyable as-is: each one's `compose.yml` pulls the orchestrator from a shared file one directory up.