Show HN: OpenBrief – Local-first video downloader/summarizer
Summary
OpenBrief is an open-source desktop app that lets users download videos, transcribe audio, generate grounded summaries, and chat with media content, all running locally on their machine.
View Cached Full Text
Cached at: 05/26/26, 12:50 AM
tantara/openbrief
Source: https://github.com/tantara/openbrief
OpenBrief
Turn videos and audio into clear, listenable briefings.
Import a video or audio file, extract its transcript, generate a grounded summary, and chat with the content — all on your desktop.
Watch the demo · Features · Model Support · Setup · Development · Roadmap
OpenBrief is a pnpm/Turborepo workspace centered on a Tauri v2 desktop app. It supports importing local media or video URLs, downloading media through bundled tools, transcribing audio, generating grounded summaries, chatting with media context, organizing playlists, and exporting reusable notes.
Features
- 📥 Import anything — paste a video link or import a local audio/video file.
- ✍️ Transcribe locally — extract captions or run on-device speech-to-text.
- 📝 Grounded summaries — generate blog-style markdown briefs with timestamped takeaways.
- 💬 Chat with media — ask questions against the summary or full transcript.
- 🔊 Listen back — turn summaries into audio with text-to-speech.
- 🔒 Open source & private — runs on your machine, free to use.
Build a searchable library from video links or local files, then extract transcripts and keep everything in one place.

Open any item to read the transcript, generate a grounded summary, and chat with the media context side by side.

Model Support
| Model type | Supported | TODO |
|---|---|---|
| Speech to text | Whisper, Parakeet, Qwen3-ASR | None |
| Text to speech | Supertonic 3, Qwen3-TTS | None |
| Large language model | OpenAI GPT, Anthropic Claude, Google Gemini, OpenRouter DeepSeek | Local Gemma 4 |
| Video embedding | None | Frame and clip embeddings for semantic search |
Repository Layout
client/
apps/
tauri/ Main OpenBrief desktop app
src/ React renderer, feature UI, domain logic, services, hooks, i18n
src-tauri/ Tauri v2 Rust boundary, commands, helper sidecar, packaging
scripts/ Helper-sidecar and media-tool preparation scripts
nextjs/ Web app and download/YouTube routes
tanstack-start/ TanStack Start app shell
expo/ React Native app shell
workers/ Worker entry points
packages/
api/ Shared API routing
auth/ Authentication integration
db/ Database schema and access
ui/ Shared UI components
validators/ Shared validation helpers
tooling/
eslint/ Shared ESLint config
github/ GitHub setup helpers
prettier/ Shared Prettier config
tailwind/ Shared Tailwind config
typescript/ Shared TypeScript config
AGENTS.md Repository development guidance
DESIGN.md Product and UI direction
Requirements
- Node.js
^22.21.0 - pnpm
11.0.9 - Rust and Cargo
- Tauri v2 platform prerequisites for your OS
Use the package manager declared in client/package.json.
Setup
Install dependencies from the workspace root:
cd client
pnpm install
If pnpm reports ignored native build scripts on a fresh machine, run pnpm approve-builds, approve the listed native/tooling packages, then rerun pnpm install.
Create local environment values when needed:
cp .env.example .env
Local Development
Use two terminals from client/ when working on both the web app and desktop app:
pnpm dev:next
The Next.js app runs at http://localhost:3000.
pnpm dev:tauri
The Tauri dev command builds the helper sidecar, starts the desktop renderer through Vite at http://localhost:1420, compiles the Rust app, and launches the desktop window.
Desktop App
Run the Tauri desktop app:
cd client
pnpm dev:tauri
Run only the renderer during frontend work:
cd client/apps/tauri
pnpm dev
Build frontend assets:
cd client/apps/tauri
pnpm build
Build or refresh bundled helper/media assets:
cd client/apps/tauri
pnpm setup:dev-sidecars
pnpm prepare:media-assets
Useful desktop checks:
cd client/apps/tauri
pnpm test:run
pnpm typecheck
cd src-tauri && cargo check
Web And Shared Workspace
Run the Next.js app:
cd client
pnpm dev:next
Run all workspace dev tasks through Turbo:
cd client
pnpm dev
Common workspace checks:
cd client
pnpm typecheck
pnpm lint
pnpm build
Database and auth helpers:
cd client
pnpm db:push
pnpm db:studio
pnpm auth:generate
Use pnpm --filter <workspace> <script> or pnpm -F <workspace> <script> for a single app or package.
Roadmap
- Improve audio file support for transcription, summaries, playback, and exports.
- Support more document and web source types, including PDFs, HTML pages, and other document formats.
- Support Parakeet ASR.
- Support Qwen3-ASR and Qwen3-ForcedAligner.
- Support Supertonic 3 TTS.
- Support local LLMs, including Gemma 4.
- Add video embedding for frame and clip semantic search across the library.
- Add voice cloning so summaries can be read aloud in a selected voice.
- Share summaries through the web and mobile apps.
- Support more artifact formats, including flashcards and other reusable study or publishing outputs.
Acknowledgements
OpenBrief builds on and takes inspiration from several projects:
- yt-dlp for video download support.
- whisper.cpp and transcribe-rs for local speech-to-text.
- FluidAudio for local Apple-platform audio AI inspiration.
- Qwen3-ASR for speech recognition model support.
- Qwen3-TTS for text-to-speech model support.
- Supertonic for Supertonic 3 TTS support.
- tweakcn for the shadcn theme.
- Voicebox and Anarlog for product and implementation inspiration.
License
OpenBrief is licensed under the GNU Affero General Public License v3.0.
Verification
Run the smallest check that proves the change, then widen as needed:
cd client/apps/tauri && pnpm test:run <pattern>
cd client/apps/tauri && pnpm typecheck
cd client/apps/tauri/src-tauri && cargo check
cd client && pnpm --filter @acme/nextjs typecheck
git diff --check
For packaging, run the relevant Tauri build on the target platform before making release claims.
Similar Articles
Show HN: VidStudio, a browser based video editor that doesn't upload your files
VidStudio is a free browser-based video editor that processes files locally without uploading, offering resize, trim, compress, watermark, subtitle and multi-track editing tools.
Show HN: Files.md – Open-source alternative to Obsidian
Files.md is a free, open-source markdown note-taking app that runs in the browser, stores files locally, and offers synchronization and a Telegram chatbot for on-the-go access.
Show HN: CPU-only transcription for YouTube, TikTok, X, Instagram videos
yapsnap is a command-line tool for transcribing video/audio from various sources (YouTube, TikTok, etc.) to plain text using only CPU, no GPU or cloud required. It leverages sherpa-onnx and yt-dlp for offline, fast transcription.
@XAMTO_AI: Guys, this is something you absolutely need but haven't discovered yet! OmniGet - one-click paste link, grab media from YouTube, Twitter, TikTok, Reddit, Udemy across all platforms. Local desktop app, no ads, no tracking. Download is super smooth. Essential for content creators, researchers, offline users. No more...
OmniGet is a free, open-source desktop app that downloads media from YouTube, TikTok, Udemy, and over 1,800 other sites without needing a terminal. It combines video, audio, course, and music downloading into one easy-to-use local application.
@shedntcare_: Someone just built the open-source tool the internet has wanted for YEARS. It’s called ReClip — and it downloads videos…
ReClip is a newly released open-source, self-hosted tool that downloads videos from YouTube, TikTok, Instagram, and 1000+ other sites, offering a clean, ad-free, and private alternative to sketchy downloader websites.