README.md

---
title: ByteBrain
emoji: 🧠
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
---

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

# ByteBrain Hugging Face App

This app generates a chalkboard-style MP4 from a topic using a simple Gradio UI.

## UI

- Logo: `bytebrain/logo/logo.png`
- Input: topic text
- Output: generated video + logs

## Output Location

All generated files are written under temporary storage:

- Linux/Hugging Face: `/tmp/bytebrain-output`
- Local Windows/macOS: system temp directory + `bytebrain-output`

You can override with:

- `PIPELINE_OUTPUT_DIR=/your/path`

## Required Environment Variables

- `GEMINI_API_KEY`
- `OPENAI_API_KEY`
- `HF_TOKEN` (recommended)

Optional:

- `HF_SPACE` (defaults to `banao-tech/vibe-voice-custom-voices`)
- `OPENAI_TRANSCRIBE_MODEL` (defaults to `gpt-4o-mini-transcribe`)
- `NARRATION_MODEL` (defaults to `gemini-2.5-pro`)

## Run Locally

```bash
pip install -r requirements.txt
playwright install chromium
python app.py
```

Open `http://localhost:7860`.

## Docker

```bash
docker build -t bytebrain-app .
docker run -p 7860:7860 \
  -e GEMINI_API_KEY=your_key \
  -e OPENAI_API_KEY=your_key \
  -e HF_TOKEN=your_hf_token \
  bytebrain-app
```

## Hugging Face Space

- Use Docker Space.
- Set the same environment secrets in Space settings.
- Entry command is already configured by `Dockerfile` (`python app.py`).