Your agent’s favorite toolfor binge-watching videos.

An MCP that connects to the AI tools you already use.

usecases

user> sort yesterday's shoot and flag the usable takes

seen · production · agency · 1:42:12

Done - I categorized all 346 takes from the shoot; 92 made the selection. Your Frame.io is updated with frame-accurate notes and tags on every take.

user> did the boom slip into frame anywhere?

seen · production · agency · 00:42

Three times. I can see it dip into the top of frame in takes 12, 31 and 77 - each one flagged as a comment right on the Frame.io timeline.

user> make our training library useful for the copilot

seen · knowledge base · 36:05

All 212 learning videos are summarized and indexed - spoken words, slides and on-screen text. Your company copilot answers from them now.

user> what did we decide about the refund policy?

seen · knowledge base · 12:41

The slide on screen reads “Refund SLA: 14 days, auto-approve under €50.” That’s the decision - not just what someone said out loud.

user> document every talk from the summit

seen · speaker event · 28:19

All 84 speakers are documented, presentation slides extracted per talk. The recap newsletter is drafted - ready to send whenever you are.

user> find the moment the keynote announced the roadmap

seen · speaker event · 14:33

14:33 - the roadmap slide is on screen: three phases, GA in March. I pulled the exact frame for your recap.

user> set up the VR capsule like in the tutorial

seen · agentic coding · 14:07

From the Details panel in the frame: Capsule Half Height 88.0, Target Bone “head.” I applied it - those values aren’t written down anywhere else.

user> which CLI flags did the deploy walkthrough use?

seen · agentic coding · 03:48

Typed on screen, never spoken: `deploy --env prod --skip-cache`. I ran the setup exactly like the video does it.

user> sort yesterday's shoot and flag the usable takes|

seen · production · agency · 1:42:12

Done - I categorized all 346 takes from the shoot; 92 made the selection. Your Frame.io is updated with frame-accurate notes and tags on every take.

See how it works ↓

Your videos contextualized, wherever they live

Drop files straight into seen, or connect Dropbox, Google Drive or Frame.io.

liveDirect upload

Drag and drop a recording. Indexed in minutes.

liveDropbox

Connect your Dropbox, browse your folders and import in bulk.

liveGoogle Drive & Meet

Connect your Drive and pull recordings from your folders. Meet recordings flow in automatically with a hotfolder.

liveFrame.io

Sync straight from your own Frame.io projects.

Got it somewhere else, a YouTube link, a webinar? Pull it down with a tool you already use, drop the file in, use the context.

Start free. Upgrade when you need more.

Plans are metered in video minutes. Team includes three seats to share the results.

Free

€0forever

30 min / month

Full pipeline: transcripts, keyframes, on-screen text detection
Entry transcription model
20 analysis minutes

Pro

€19/ month

300 min / month

Everything in Free
Standard transcription model
API access
120 analysis minutes

Team

€79/ month

2000 min / month

Everything in Pro
3 seats included · shared workspace & comments
500 analysis minutes

Pro transcription (our best model) is available on every plan, paid per video in analysis minutes.

What teams use seen for

seen remembers not just what was said, but what was shown and what happened, down to the exact second.

Every call remembered, transcripts free

Record any Meeting call with one click
Action items, chapters and who said what, automatically
Presentations detected and ready for deep visual analysis

From a video only humans could watch,to knowledge an agent can use.

Upload your videos and seen makes everything inside them searchable, from spoken words and on-screen text to the exact moment something appeared. Computer vision has never seen such possibilities.

→

transcript · 14:02

“…scale the capsule to the player’s head height, then rebind the IK targets…”

on-screen text · 14:07

Details panel: Capsule Half Height 88.0 · Target Bone: head

frame caption · 14:07

Blueprint graph wiring a Scale-By-Head-Height node into the IK setup.

summary · topic · tools

VR body IK calibration in Unreal Engine 5. UE5, VR Expansion Plugin

Live in three steps

Connect an agent, fill it with your recordings, and open the knowledge up to your tools and your team.

Connect

One click adds seen to Claude, Cursor or Codex.

Fill it

Drop in recordings or connect a source. Transcript, keyframes and on-screen text are indexed in minutes, and it's instantly queryable.

Open it up

Give agents API access to the content, or share transcripts and clips with your team. Your knowledge, on tap.

Your agent finished the whole seasonbefore you even started it

More than what was said

seen understands spoken words, on-screen text, interfaces, slides, diagrams, and what actually happens in the video.

Everything stays connected to the video

Every piece of information leads back to the exact moment and frame it came from. The original context is always one click away.

Find anything. However you remember it.

Ask a question, describe a scene, or search for something you saw. You don’t need the filename, timestamp, or exact words.

Turn recordings into working knowledge

Make tutorials, meetings, product demos, research, lectures, and entire video archives instantly useful again.

Available to every agent

Connect seen once and let ChatGPT, Claude, Cursor, Codex, or your own agents query the same video knowledge - never trapped in one system.

Private by design - your videos stay in your workspace, are never used for training, and can be deleted anytime.

Stop re-watching. Start working.

Talk to us

Good to know

Can agents change my library, or only read it?

Both. Over MCP and the REST API your agents search the content and reorganize it (tag, group into playlists, import) acting on your workspace directly.

Who can see my videos? Do you train on them?

No training, ever. Every workspace is isolated to your own uploads, no public videos, no scraping, no cross-tenant access.

What can I put in, and how is it billed?

Common video formats, short clips to long recordings. You’re billed by the minutes you index, not per user, Team includes three seats.

What’s beyond direct upload?

Connect today with Dropbox, Google Drive or Frame.io - meeting recordings auto-file into a Meetings playlist. Enterprise adds SSO, a DPA, dedicated processing and regional storage.