AI video composition
Ask for the edit.
Describe the story, message or argument you need. Anjin Media finds the right moments in your video and composes a finished edit to your chosen length and format.
One video or an entire archive. Available through the platform, API and MCP.
The brief
“Find and compose the strongest edit for social. Choose the moments, structure and pacing most likely to hold attention.”
Edit ready to review
One instruction sets the story, duration and format. Anjin finds the moments, builds the cut and returns the rendered edit.
Make your first editFinished edit · press play
Inside the editSee how five source moments became one story
Source
Jocko Willink, Matt McCusker & Jeff Dye, Mostly Wise #3 ↗The cut opens on the payoff, rewinds into the encounter, raises the stakes with two surprising details and closes on the eight-foot jump. Repeated setup was left out to keep the story moving.
- Hook · 28:05–28:20 · Source AThe visceral payoff: the dog launches at full speed.
- Body · 27:21–27:35 · Source AResets to the first encounter and builds anticipation.
- Body · 27:44–27:54 · Source AThe quiet ‘locked in’ beat raises the tension.
- Body · 18:26–18:46 · Source AA vivid warning supplies the most surprising detail.
- Close · 27:07–27:16 · Source AThe eight-foot jump ends on a memorable image.
Brief in. Finished edit out.
Upload and index once
Your footage is transcribed with word-level timecodes and indexed. Every moment becomes addressable - this brief and every one after it.
Send the brief
Subject, audience, duration, aspect - written the way you would brief an editor. In the app, over the API, or from an agent through MCP.
Receive the edit
Rendered, on-brand, at the length and format you asked for - every segment traceable to its source timecodes.
One archive. Many finished stories.
Different voices, formats and ideas, composed from long-form footage with a brief. Select any edit to watch.
Select an edit to watch
More than a clip.
Clip tools scan your footage for a spike and hand back an isolated moment. A composition executes an editorial brief.
A clipper returns
One highlight, out of context.
A single stretch of tape, chosen because it spiked - not because it says what you need said. Making it mean something is still your job.
Anjin Media returns
One coherent narrative.
Several relevant moments, retrieved from across your footage - repetition removed, ordered to carry the argument, grounded entirely in what was actually said.
Five things this is built on.
Compose, not clip.
Multiple segments, selected and ordered to carry one argument - not a single moment cut out of context.
Your archive becomes useful.
Index footage once and every word of it stays addressable - ready to answer the next brief, and the one after that.
Built for agents.
The same composition capability over REST and MCP, so your tools and agents can ask for the edit too.
Grounded in your content.
Every segment carries its source timecodes and transcript - provenance you can check, moment by moment.
Brief to finished edit.
Set duration and aspect, review the cut plan, receive a rendered edit - not a timeline of suggestions.
Your archive has more to say.
Years of interviews, webinars and productions are sitting in storage, remembered by the few people who were there. Connect them once, ask across the recordings a brief names - up to four at a time - and compose new work from what they hold, every edit grounded in footage you already own.
The archive stops being a storage cost and starts answering briefs.
Give your agent video-editing capability.
One request: the brief, the sources, the formats. Back comes a composition job whose finished edit carries its cut plan, source timecodes and transcript. The same capability is exposed over MCP, so agents can compose edits as a step in their own work.
POST /v1/cuts
{
"prompt": "60 seconds on why we changed our pricing - the clearest
explanation, for LinkedIn",
"source_group_ids": ["9f4b…"],
"aspects": ["9:16"],
"auto_render": true
}
→ { "id": "c7d2…", "status": "planning" }Pay for the footage you bring in.
Plans meter source minutes once, at ingest. Searching, briefing, revising and re-rendering are free.
Creator
£39/month
- 240 source minutes a month (4 hours of footage)
- Single camera
Pro
£99/month
- 600 source minutes a month (10 hours of footage)
- Single camera
Studio
£249/month
- 1,500 source minutes a month (25 hours of single-camera footage)
- Two cameras, each metered
Enterprise from £1,500/month - persistent archive, contracted compositions.
Questions
- What does Anjin Media actually return?
- A finished, rendered edit - and the cut plan behind it. Every segment lists its source video, transcript text and source timecodes, so you can review the plan before rendering and verify each moment after.
- How is this different from an AI clipper?
- A clipper isolates one highlight at a time. Anjin Media executes a brief: it selects several relevant moments - they don't need to be adjacent - removes repetition and composes them into one narrative at the duration and aspect ratio you set.
- What footage can I use?
- Your own. Upload long-form video - interviews, webinars, talks, productions - and it is transcribed with word-level timecodes and indexed once. From then on, every moment in it is addressable from a brief.
- Can my AI agent drive it?
- Yes. The full workflow - sources in, brief in, edit out - is available through a REST API and an MCP server, so agents and tools can create compositions the same way a person does in the app.
- What does it cost?
- Paid plans start at £39 a month for 240 source minutes - four hours of footage. Minutes are counted once, at ingest; searching, briefing, revising and re-rendering cost nothing. Enterprise archive plans start at £1,500 a month.