4 commits

Author	SHA1	Message	Date
Natalie	3378d0710d	feat(broadcast): backend-api + controller + shared refinements from quinn-platform-architect agent (pure Bun.serve no-dep backend, robust dest store with /tmp fallback + atomic, typed surfaces, core delegation, error handling; no hono dep for immediate verifiability) ... Scoped to agent's paths. Self-verified per agent's commands (syntax, store, boots). Per approved plan + multi-agent on main.	2026-06-28 14:44:55 -04:00
Natalie	09346e536e	feat(broadcast): multi-agent parallel delivery of full feature on main -- backend-api, frontend-public, shared, controller refactor into modules (llm-agent, obs-client, etc.), mcp-server start, deployment surface quinn.cast, obs scene collection (per approved plan + user 'merge to main + use multiple agents')	2026-06-28 14:38:01 -04:00
Natalie	8a1bbd7f29	fix(broadcast): make /data dir creation robust in controller (no spurious warning on first run)	2026-06-28 13:30:49 -04:00
Natalie	f49b7df43d	feat(broadcast): LLM-controlled remote OBS + SRT relay for low-upload hotel streams with high-bitrate multi-RTMP from DO ... - New scripts/provision-stream-droplet.sh (modeled on gpu one) for cheap DO droplet with docker + v4l2loopback + mediamtx ready. - scripts/hotel-srt-push.sh: macOS ffmpeg one-liner helper for reliable modest-bitrate SRT contribution over bad WiFi. - New @features/broadcast/controller/: self-contained Bun server with embedded Tailwind chat UI, full xAI Grok-4.3 tool-calling agent, pure-WebSocket obs-websocket v5 client, dynamic destination + ffmpeg fanout manager. - Full docs + RUNBOOK with architecture, provisioning, hotel push, OBS notes, security, troubleshooting, and verification notes. - Matches the 'RTMP relay for simulcast' vision already promised in performer marketing. The hotel side sends ~3 Mbps SRT. All compute, final encode, and broadcast bitrate (to Twitch/YouTube/etc.) happens on the DO network. Control is natural language chat to the LLM.	2026-06-28 13:30:39 -04:00