AI screen recording agent for autonomous, remote video creation

Auto Capture is Tella's agentic screen recorder, a camera for browser agents. An AI agent opens a web page in its own browser, explores it, plans a walkthrough, and records it with an animated cursor, with no one at the keyboard. You get back a Tella recording you can edit, narrate and share like any other.

Agentic screen recording means an AI agent makes the recording for you. Instead of you clicking through a product while a screen recorder runs, you describe the demo you want and the agent explores the page, decides what to show, checks its script works, then films it. Tella's Auto Capture does this for web pages.

Yes. Auto Capture is built to be called by agents. Claude, ChatGPT, Claude Code, Codex, Cursor or any app connected to Tella's MCP server at https://api.tella.com/mcp can start a take with create_auto_capture, poll it with get_auto_capture, and place the finished recording in a Tella video with create_video. The Tella Agent in Tella's editor can use it too.

Connect Claude to Tella's MCP server, then ask it for a demo with your product's full URL and who it's for: "Use Tella to record a product demo of https://example.com for first-time visitors." Claude starts an Auto Capture take, checks on it until it's delivered, and puts it in a new Tella video, ready to trim, caption and share.

Ask the Tella Agent in the editor, or an AI app such as Claude or ChatGPT connected to Tella's MCP server, for a recording of a page with its full URL. For example: "Record a 30-second tour of https://example.com/pricing for first-time visitors. End on the yearly toggle."

Yes. Any agent connected to Tella's MCP server can start a take, including Claude Code or Codex running in a CI job or a scheduled automation, using a Tella API key as the bearer token. Point it at a URL the capture browser can reach: a preview deploy on a public domain works, localhost doesn't. If the preview needs a sign-in, save a login under Settings, then Credentials.

Yes. Save a login for your product under Settings, then Credentials. The agent signs in off camera while it explores, and the recording starts already signed in. Password fields are always masked. Pick the login from 1Password and Tella fills in its one-time code too; otherwise, if your product sends a sign-in code, you enter it on a Tella page, never in the chat.

It records http and https pages on public domains. It can't record localhost, IP addresses or intranet hosts, desktop or mobile apps, or sites that block automated browsers with a "verify you're human" check.

A screen recording with an animated cursor, plus a handoff note on what it shows, chapter markers, and suggested cuts for page loads and long pauses. Edit it like any other clip: trim it, add zooms, restyle the cursor, add a voiceover or change the background.

Playwright and Puppeteer film a script you write and give you a raw browser video. Auto Capture writes and tests the script itself from a plain-language request, films it with an animated cursor, and delivers it into Tella's editor with chapters and suggested cuts. No code, and no script to update when your UI changes: ask for a new take.

Yes, for web apps. Tella's Auto Capture records the walkthrough in its own browser, and the Tella Agent, Claude or ChatGPT then edits it into a finished video: cuts, zooms, captions and chapters. Tella doesn't generate AI narration, so add a voiceover you recorded or record narrated parts yourself. For desktop apps, record your screen with Tella's recorder and let the agent do the editing.

No. Tella doesn't generate AI narration. Add a voiceover you recorded elsewhere as a sound in the editor, or record yourself narrating in Tella and use the Auto Capture take as b-roll over that part.

Auto Capture is in beta and included with paid plans. It uses your workspace's credits, the same balance the Tella Agent spends, charged when each take ends.