90210 – running the show without property tax

Hacker News Top Tools

Summary

90210 is a production-grade local app that uses AI models like Google Veo 3.1, Gemini 2.5 Pro, and ElevenLabs Music to turn screenplays into finished short films with synchronized video, audio, dialogue, music, and subtitles, featuring a quality oracle for auto re-rolls.

No content available
Original Article
View Cached Full Text

Cached at: 06/08/26, 03:15 AM

Achint08/90210

Source: https://github.com/Achint08/90210

📍🎬 90210

Like the show. But you’re the showrunner.

All the drama. None of the property tax.

A production-grade local app that turns a screenplay into a finished short film with synchronized video, native audio, dialogue, music, and subtitles. Powered by Google Veo 3.1, Nano Banana, Gemini 2.5 Pro, and ElevenLabs Music.

Every quality decision is mathematically grounded: a composite Quality Oracle (VideoScore, ViCLIP, LAION aesthetic, MUSIQ, ArcFace, LPIPS, Whisper) drives auto re-roll and adaptive tier escalation. A Story Oracle (Reagan-6 arcs, TRIPOD turning points, Average Shot Length, SBERT coherence) keeps the narrative on target.

Quick start

# 0. one-time
python3 -m venv .venv
.venv/bin/pip install -e ./backend          # installs FastAPI + Pydantic + Google + ElevenLabs deps
cp .env.example .env                         # fill in GEMINI_API_KEY and ELEVENLABS_API_KEY

# 1. Backend
./run-backend.sh                             # starts uvicorn at http://localhost:8000
# (the script sets PYTHONPATH because Homebrew Python 3.14 ships its own
#  sitecustomize.py that breaks venv editable installs — PYTHONPATH sidesteps it.)

# 2. Frontend (new shell)
cd frontend
pnpm install                                 # or npm install
pnpm dev                                     # → http://localhost:3000

# 3. iPhone access
#    - same WiFi: http://<laptop-ip>:3000 → Safari Share → Add to Home Screen
#    - off-WiFi: tailscale serve / funnel (see docs)

# 4. Optional Quality Oracle (heavy ML deps; ~3GB)
.venv/bin/pip install -e "./backend[oracles]"

Documentation

  • Design spec: docs/superpowers/specs/2026-05-09-screenplay-to-movie-design.md
  • Veo 3.1 reference: docs/superpowers/research/veo-3.1-deep-dive.md
  • Cost / latency / quality math: docs/superpowers/research/cost_latency_quality_math.py

Required API keys

ServiceUsed forCost (May 2026)
Gemini APIVeo 3.1 (video+audio), Nano Banana (portraits), Gemini 2.5 Pro (Director)$0.10–0.40/sec video
ElevenLabsMusic tracks$0.01/sec music

A 2-min movie typically costs ~$20.

Architecture

See spec §4. Local FastAPI backend + Next.js PWA frontend. PWA installs to home screen on iOS Safari for native-app-like experience.

Similar Articles

@IndieDevHailey: HKU open-sources ViMax — AI one-person film crew, one-click to produce a complete long-form video! One-sentence idea → automatically generates a finished production with script, storyboard, voiceover, and consistent characters! Solves these pain points: - No more short clips of a few seconds where characters change faces - Multi-agent collaboration: screenwriter writes story, director controls shots, producer ensures consistency, generator outputs video...

X AI KOLs Timeline

HKU open-sourced ViMax, a multi-agent collaborative long video generation tool that can generate a coherent video with script, storyboard, voiceover, and consistent characters from a single sentence, solving issues like short video fragmentation and character inconsistency. The developer also introduced Taste-Skill, a frontend framework that improves the aesthetics of AI-generated interfaces.