Web UI
How to use the OpenReels web interface to create videos, watch the pipeline, and download results.
Starting the web UI
The web UI runs via Docker Compose. It starts three services: Redis (job queue), the API server, and a background worker.
docker compose upOpen http://localhost:3000 in your browser.
Submitting a job
The home page presents a topic input with provider controls.
1. Enter a topic
Type any topic into the input field. This is the subject of your video. Examples:
- "The Apollo 13 disaster, from explosion to miraculous return"
- "5 stoic lessons that changed my life"
- "How coffee changed the course of history"
You can also click one of the suggested topics from the inspiration section below the input. Topics are grouped by category (History, Science, Culture, Technology) and can be shuffled.
2. Choose a visual style
Below the input, a horizontal gallery shows all 14 visual archetypes. Click one to lock in that style, or leave it on Auto Style to let the LLM choose the best fit for your topic.
3. Configure providers
The input area includes quick controls for Platform (YouTube, TikTok, Instagram) and Pacing (Fast, Moderate, Cinematic, or Auto).
Click Advanced to expand the full provider panel:
- LLM Provider — Anthropic, OpenAI, or Gemini
- TTS Provider — ElevenLabs, Inworld, Kokoro, Gemini TTS, or OpenAI TTS
- Image Provider — Gemini Imagen or OpenAI DALL-E
- Music Provider — Bundled (free) or Lyria 3 Pro ($0.08/track)
- Style Override — Same as the archetype gallery, as a dropdown
- Dry Run — Toggle on to output the DirectorScore JSON without generating assets or spending money
- Creative Direction — Free-form textarea where you can describe your creative vision: visual style, mood, script notes, scene ideas, music preferences. The AI reads it like a brief from a producer. A character counter shows usage against the 10,000 character limit.
- Replay from Score — Upload a previous
score.jsonfile to skip research and director stages. Useful when a previous run produced a great creative plan but downstream stages failed.
4. Generate
Click Generate to submit the job. You are redirected to the job page.
Watching the pipeline
The job page shows each pipeline stage as a card:
| Stage | What you see |
|---|---|
| Research | Live research data as the AI gathers facts via web search |
| Director | The DirectorScore with script, scene breakdowns, and visual direction |
| TTS | Voiceover generation progress |
| Visuals | Scene-by-scene asset generation. AI images, AI video clips, and stock footage appear as they are produced. Failed assets show inline errors |
| Assembly | Remotion render progress with frame count |
| Critic | AI quality review with scores. If quality is below threshold, the pipeline automatically re-runs |
All stages stream progress via Server-Sent Events (SSE). You see each stage transition from pending to active to complete in real time. A cost breakdown card shows estimated spend across LLM, TTS, images, video, music, and stock.
The storyboard panel displays the full scene layout with generated assets once the Director stage completes.
Downloading the result
When the pipeline completes successfully, the job page shows a Completed panel with a download button. Click it to download the rendered MP4 file.
If the job fails, the page shows an error panel with the failure reason. You can return to the home page and try again with different settings.
Gallery
Navigate to the Gallery page to browse all previously generated videos. The gallery supports search, filtering, and bulk deletion. Each card shows the video thumbnail, topic, archetype, and creation date.