Scrib listens, transcribes with speaker labels, and turns conversation into action items — automatically. Your AI agent then handles the follow-ups.
Three things, done well: capture the conversation, understand it, and act on it.
Scrib runs real-time speech-to-text with diarization. You see who said what, as they say it — across phone calls, in-person meetings, and video conferences.
When the meeting ends, your AI reads the transcript, pulls out tasks, identifies who they're for, and surfaces dates. Review, edit, dismiss — or just let it run.
Connect Google, Microsoft, Slack, and GitHub. The agent runs on the server, so you can fire off commands by text or voice — "schedule the QA review for Wednesday at 10" — and it's done.
Hit record in the browser or app. Scrib listens and transcribes in real time.
Run your meeting like normal. We label speakers and track context.
Tap stop. Get a summary, transcript, and a list of action items in seconds.
The agent schedules, messages, and follows up across your connected tools.
Connect once in your dashboard. Your agent handles the rest.
Free during the beta. $5/month after launch.
Your transcripts and audio live on the Scrib server. We never train AI models on your data. You can export or delete everything anytime from your dashboard.
You can use ours during the beta. Once we launch, Pro users supply their own Anthropic or OpenRouter key — you only pay for what you use, on top of the $5/month subscription.
Yes — record from the mobile app while on a call, or upload the audio after. In-person meetings work the same way: open the app and hit record.
We use sortformer-based diarization, which performs well even with overlapping speech. Accuracy improves the more you use it — speakers are remembered across sessions.
Yes. One click in the billing section. No phone calls, no emails, no exit interviews.
We're letting people in week by week. Drop your email and we'll send an invite.