← Back to projects
Briefly AI
End-to-end processing pipeline: download/transcode via yt-dlp + FFmpeg, transcription with Whisper, and LLM summarization. Includes JWT auth, job history, strict prod setup (Docker + health endpoints), and a strong test suite (Vitest + Playwright) with full mock mode for cheap and fast E2E runs.
Next.jsReactTypeScriptMongoDBMongooseTailwind CSSServer-Sent Events (SSE)OpenAI (Whisper + LLM)FFmpegyt-dlpVitestPlaywrightDocker
Screenshots
UI snapshots showing the main flows and features. Click to enlarge.