What is Sumarii? Recording to usable notes

After a lecture, podcast, call, or meeting, someone still has to turn the recording into something readable. That work usually falls on a person, even when you already have a transcript.

Transcription is not the deliverable

Most tools stop at speech-to-text. That step matters, but it is not what you share or reuse.

A useful output needs structure that matches the content: key points for a lecture, highlights for a podcast, decisions for a team call, pull quotes for an interview. Different inputs need different templates and different things extracted.

Sumarii is built around that structured output, not a wall of raw text.

What Sumarii does

One session, four views:

1. Input: Live mic recording, audio/video file upload, or direct media URL
2. Transcript: Speaker diarization, word-level timestamps, playback sync, inline review of low-confidence words
3. Notes: AI summarizes the transcript into templated notes (meetings, lectures, podcasts, interviews, and more). Edit markdown inline or regenerate
4. Insights + chat: Talk time and pace computed in the browser; AI chat grounded in the session to ask what was said or refine the notes

Transcription is infrastructure. The product is recording → transcript → notes you can edit and export.

Who it is for today

Anyone working with spoken audio or video:

• Students and teachers turning lectures into study notes
• Podcasters and creators who want show notes or episode summaries
• Teams that need structured notes from calls and workshops
• Researchers and journalists reviewing interviews

We are early. One account gets one personal workspace. Shared team workspaces are not available yet.

What you can use now

• Browser app at app.sumarii.com
• Windows desktop app for live recording (mic + system audio); macOS planned
• Templates for different session types; custom templates in your workspace
• Export docx, txt, srt, and audio (PDF not available yet)
• Sign in with Google or email and password

Alpha access

Sumarii is in alpha testing. There is no checkout today. Create an account and use the product free with generous limits (including about 6 hours of transcription per month while we tune the product).

We read feedback and bug reports. If something breaks or feels wrong, use the help form on the site.

Details: alpha access page.

Pricing (planned, not active yet)

We are preparing Free and Pro plans (and team workspaces later). Paid tiers and limits are defined in the product but billing is not turned on during alpha. Everyone uses the alpha pool for now.

When we open payments, we will announce it on the site before charging anyone. Questions about future pricing: contact us.

Security

• Meeting and session content is not used to train our models or provider models we use for inference
• Customer data is stored on infrastructure in Japan, in principle
• HTTPS in transit; encrypted at rest

More detail: security & trust · privacy policy.

Try it free. Run one recording and see transcription through to finished notes.