Sea Island Cotton AI

Spoken words, put to work.

We build AI products that capture what people say and turn it into notes, follow-ups, study material and clean text for your own AI models. Three of them are in market today.

See the products Talk to us

What we make

Three products, one transcription engine

The same ActaScribe engine runs under all three. Use it in an app, call it from your own code, or set it on the table in your conference room.

ActaScribe app

Record it once. Get the notes, the follow-ups and the full transcript.

Record a meeting on your phone, upload audio, paste a YouTube link, or turn on Stage Mode during a live talk. Every recording comes back in the same structure: a summary, main points, action items, follow-ups, stories, references, counterarguments and the full transcript. You can customize that template for each type of meeting.

ActaScribe never uses your conversations to train AI, and default retention is zero. Pro starts at $19 a month, billed annually, with a 7-day trial.

Start a 7-day trial Visit actascribe.ai

Set up for the way you work

Speakers

Stage Mode puts live captions, an audience word cloud and Q&A on screen while you talk. TalkBase turns your talks into a chatbot that answers with citations.

Coaches

Build a chatbot from your recorded sessions, so learners can get answers between calls without waiting on you.

Students

Turn lectures and YouTube videos into structured study material you can actually review.

Sales

Record the meeting, then send the prospect a summary with the action items you both agreed on.

ActaScribe API

Clean transcripts, ready for your own models

Send a URL and the API does the rest. It fetches the audio, transcribes it and returns plain text with timed segments that you can load into a custom model, a search index or a training set. It accepts a direct audio link, a single YouTube video, or a YouTube playlist of up to 100 videos.

REST and JSON at api.actascribe.ai/v1
Bearer keys with scoped permissions, optional expiry dates and IP allowlists
Asynchronous jobs. A short audio file usually finishes in under a minute.

Read the API docs Ask about volume

Request

curl -X POST https://api.actascribe.ai/v1/transcripts \
  -H "Authorization: Bearer as_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "youtube_url": "https://youtu.be/...",
    "language": "en"
  }'

202 Accepted

{
  "id": "5b8a3f1c-7d2e-...",
  "status": "pending"
}
// then GET /v1/transcripts/{id}

ActaScribe Capture Puck

Meeting transcription that stays in your building

Set the puck in the middle of the table and plug it in. Four microphones pick up voices from every direction, up to 5 meters (16 feet) away, and the puck's own eight-core computer transcribes as people talk. When the meeting ends you get a final transcript with speaker labels, plus a summary and action items. Your audio never goes to a public cloud or an outside AI company.

We built it for conversations that can't go to a cloud note-taker: attorney-client meetings, clinical and counseling sessions, board meetings, HR discussions and pastoral care.

StandaloneEverything runs on the puck. Audio never leaves the device.

HostedThe final pass runs on 6 Levers AI's private servers in a U.S. facility, over an encrypted tunnel.

On-siteThe final pass runs on your own Prediction Guard server. Audio never leaves your building.

360° pickupFanless and silentEncrypted storageSame API as cloud ActaScribe

Talk to us about the puck Explore it in 3D

Contact

Tell us what you're working on

Send a note and someone from our team will be in touch soon. We'll also email you a copy, so you know it arrived.