Tella Agent

The AI agent that edits your video

Say what you want. Watch every cut, zoom and layout land.

Ask for the edit, not the steps

The agent reads your transcript and timeline, makes each change in the editor while you watch, and undoes or redoes anything you ask.

Cut and tighten

"Cut the filler words and the long pauses."

Zooms and layouts

"Zoom in when I click Settings. Split layout for the demo, full camera for the intro."

Subtitles and chapters

"Add subtitles and translate them to Spanish. Write a title and chapters."

Record missing footage

With Auto Capture (closed beta), the agent can film a web page your video needs and place it as b-roll.

Real edits on your real recording

Agentic editing isn't video generation. The agent uses Tella's own editing tools on footage you recorded, so you keep your voice, your screen and your timeline, and can take over at any point.
What is agentic video editing?

Built to learn how you edit

Point at the transcript

Select words in the transcript and add them to the chat to edit exactly that part.

Memory

Tell it your preferences once, like your brand color, for you or the whole workspace.

Skills

Save an edit you repeat as a skill, or install one from GitHub, and run it with /.

Pick the model

Choose the AI model and how hard it thinks. Bigger models handle harder edits.

Or bring your own agent

Prefer Claude, ChatGPT, Codex or Cursor? Connect them to Tella's MCP server for more than 100 editing, export and publishing tools, on every plan.
Connect your agent

Frequently asked questions

The Tella Agent is an AI video editing agent built into Tella's editor. Describe the change you want in plain language and it edits your video step by step while you watch: cuts, layouts, zooms, subtitles and more. It's the chat in the sidebar on the right of the editor.
It works with the same tools you use in the editor. Ask it to cut filler words and long pauses, remove a section, zoom in when you click a button, use a split layout for the demo, add and translate subtitles, or write a title and chapters. With Auto Capture, in closed beta, it can also record a web page and place the footage in your video.
No. It edits the video you recorded, with real editing tools, so it never invents what you said or did. It can add generated images or sound effects where you ask for them.
No. The Tella Agent is in public beta and is in the editor for everyone. Sending it messages needs a paid plan.
It's included with paid plans (Pro, Premium and Enterprise) and runs on credits. Each Pro seat includes 3,000 credits a month and each Premium seat 6,000, pooled across the workspace. During a free trial you can edit with AI from Claude or ChatGPT through Tella's MCP server instead.
Yes. Connect Claude, ChatGPT, Codex or Cursor to Tella's MCP server and they get the same kind of editing tools. The MCP server is included on every plan, including the free trial.
Yes. It remembers preferences you tell it, for you or for the whole workspace, and you can save a multi-step edit as a skill that runs the same way every time.
Tella does. Ask the Tella Agent to "cut the filler words and long pauses", or use Auto Cut in the editor, which also removes mistakes and restarts. Claude or ChatGPT connected to Tella's MCP server can do the same with its remove_fillers and remove_silences tools.
Tella. Record your screen and camera, then ask the Tella Agent for the polish: cut mistakes and pauses, zoom in when you click, switch between screen and camera layouts, add captions, a title and chapters. It edits your real recording step by step while you watch, and you can take over at any point.
Record the demo in Tella, then ask the Tella Agent: "Add captions, my logo, my brand color for text, and zoom in on my clicks." Tell it your brand preferences once and it remembers them for you or the whole workspace. Save the edit as a skill and run it on every new demo with /.
Record in Tella and describe the result you want to the Tella Agent in plain language, for example "make this a tight two-minute how-to with zooms and captions". It plans and makes the edits, and you can ask it to undo or redo anything. Nothing to learn about timelines or keyframes.
Tella. Zooms can follow your clicks automatically, and the Tella Agent adds them on request ("zoom in when I click Settings"). Recordings made with Tella's Mac or Windows app also get an animated cursor you can restyle, smooth and add click effects to.
No. Tella doesn't generate AI narration. Record your voice with your screen, re-record a clip to redo its narration, or add a voiceover you recorded elsewhere as a sound.

Record a video and let the agent edit it

Get started