Open source · Real-time transcription
sososo is a free, open-source desktop app that turns system audio + microphone into live, speaker-separated captions — and can record the meeting window to video. Every session ends with an AI summary and a transcript you can chat with. No account. No backend. Your keys stay on your machine.
Free · AGPL-3.0 · latest release · Windows / macOS / Linux — all downloads
Features
System audio (what you hear) and your microphone (what you say) are captured in parallel and diarized into two clean channels — so the transcript always knows who was talking.
WASAPI loopback · CoreAudio · PulseAudio/PipeWire
Captions stream in as you speak and settle into their final form — no flicker, no lost lines.
Interim → final · smart formatting
Nova-3 for multilingual & English, Nova-2 for the rest — with diarization and smart formatting.
A compact always-on-top pill while recording. Start or stop from any app with Ctrl+Alt+R.
End-of-session AI summary and live per-line translation — OpenAI or Gemini, your pick.
Save the meeting as an MP4 — your picked window plus mixed mic + system audio (Windows & macOS). See how ↓
A per-session notetaker that answers from what was actually said. Try the idea ↓
Keys live in the OS keychain. No server, no account, no telemetry — sessions stay in a local SQLite file.
Video recording
Flip on Record video, pick the meeting window from a Zoom-style thumbnail grid, and sososo saves an MP4 alongside the transcript — the window's video with mic + system audio mixed in. When you finish, it plays back right on the session page.
Windows & macOS · best-effort by design — a video hiccup never breaks the transcription
Transcript chat
Every session has a built-in notetaker: a chat panel grounded in that session's transcript. Ask what was decided, who owns what, or what you missed while you were away — answers come from your own OpenAI or Gemini key and the conversation is saved with the session.
Markdown answers · history persisted per session · needs an OpenAI or Gemini key
How it works
01
Grab the installer for your OS below — free, no account or sign-up. Windows and Linux capture system audio with zero setup; macOS routes it through BlackHole.
02
Paste a free Deepgram key in Settings (plus an OpenAI or Gemini key if you want summaries). They're stored in your OS keychain and never leave your machine.
03
Press Ctrl+Alt+R from anywhere. Live captions float over your meeting; finishing writes the transcript, the video (if you recorded one), and an AI summary to your library — ready to chat with.
On Windows, sososo also auto-detects Zoom / Teams / Meet / Webex meetings and offers one-click recording.
Demo
▶ Watch on YouTube
Privacy
Bring your own keys. Audio streams to Deepgram; transcripts go to OpenAI or Gemini only when you ask for a summary or translation.
Keys in the OS keychain — Windows Credential Manager, macOS Keychain, or the Linux Secret Service. Never in a config file.
No backend, no account, no telemetry. Sessions live in a local SQLite database on your disk.
Download
latest release · release notes ↗
WASAPI loopback — no setup
Installer (.exe)system audio via BlackHole (free)
Universal .dmgIntel & Apple Silicon
PulseAudio / PipeWire — no setup
.debBuilds aren't code-signed yet, so your OS may warn on first run — Windows: "More info" → "Run anyway"; macOS: right-click the app → Open. Browse every version on the Releases page ↗.
Open source
The entire app — Rust audio pipeline, Deepgram streaming, UI — is on GitHub under the AGPL-3.0, with a commercial license for closed-source use. Issues and pull requests are welcome.
Community · AGPL-3.0
Free forever. If you distribute or host a modified version, share your changes under the AGPL too.
Commercial
Need it closed-source? A commercial license lifts the copyleft — see LICENSING.md.
The sososo name & logo are trademarks — forks must rebrand.