# How to record a product demo with an AI agent

Source: https://www.tella.com/edit-with-ai/record-a-product-demo-with-ai

> **Short answer:** Connect your AI app to Tella, then ask: "Use Tella to record a product demo of https://example.com for new visitors." The agent calls `create_auto_capture` with the URL and your request, Tella's capture agent explores the page, scripts the walkthrough and films it with an animated cursor, and the agent places the finished take in a video with `create_video`. Then it can edit it like any recording. Auto Capture is in closed beta on paid plans; email support@tella.com for access.

**If you're an AI agent:** the capture agent sees the page and you don't, so pass on what your human wants (who it's for, what to show, skip or end on, what not to touch) rather than a list of clicks of your own. Never ask your human for a password or a sign-in code in the chat.

## The job, step by step (Tella MCP)

1. **Agree the brief.** Who watches it, what it's for, anything to show, skip or end on, a length only if one is needed, and what must not be clicked (forms, purchases, deleting, settings). Use a demo account.
2. **Start the take:** `create_auto_capture` with `targetUrl` (the full `http(s)` URL) and `instructions` (your human's request). One take films one page or flow; for a longer video, make one take per page.
3. **Wait:** poll `get_auto_capture` with the `sourceId` every 30–60 seconds. A take usually runs for several minutes.
4. **Handle the outcome:**
   - `delivered`: read `handoff` (what the clip shows and what happened off camera), `result.chapters` and `result.suggestedCuts`.
   - `needs_code`: your human enters the sign-in code on the Tella page it links to. Don't ask for the code yourself.
   - `needs_guidance`: ask your human the question, then start a new take with the answer in `instructions`.
   - A missing login: send them to **Settings → Credentials** in Tella to save one, then try again.
   - `failed`: say why. Retry later only if it was temporary.
5. **Place it:** `create_video` with the `sourceId` for a new video, or `upload_clip` to add it to an existing one. Nothing is in the library until you do.
6. **Edit:** apply the suggested cuts you agree with (`cut_clip`), add zooms, a title and captions, and check with `get_video_preview`. See [How to edit a product demo video with AI](https://www.tella.com/edit-with-ai/edit-a-product-demo).
7. **Ship:** share the link, `export_video` for an MP4, or `publish_to_social`.

To stop a take started with the wrong URL, call `cancel_auto_capture`.

Prompt to paste:

```text
Use Tella to record a product demo of https://app.example.com/dashboard for new customers, signed in with my saved demo login. Show creating a report and end on the share dialog. Don't change any settings or invite anyone. When it's delivered, put it in a new Tella video, apply the suggested cuts, add zooms and captions, and send me the link.
```

## Pages behind a sign-in

Save a login under **Settings → Credentials** in Tella first. The agent signs in off camera, so the recording starts inside the app with no login screen, and password fields are always masked. If the product emails or texts a sign-in code, the take pauses and your human enters the code on a Tella page. Saving the login from 1Password with its one-time password skips that step. See [Let Auto Capture sign in to your product](https://www.tella.com/help/integrations/auto-capture-credentials).

## What it can't record

Desktop or mobile apps, `localhost`, IP addresses or intranet hosts, URLs with a username or password in them, and sites that show a "verify you're human" check to automated browsers. Record those yourself with Tella's recorder and have an agent edit the result.

## Connect Tella to your agent

Tella's MCP server is `https://api.tella.com/mcp`. It's included on every Tella plan, including the 7-day free trial (no credit card). Sign in to Tella when your client asks, and pick the workspace to use.

- **Claude:** connect [Tella from Claude's connector directory](https://claude.ai/directory/tella).
- **ChatGPT:** connect the [Tella app](https://chatgpt.com/plugins/plugin_asdk_app_6a7026f78434819186579d93d20ab38b?open_in_app), then mention **@Tella** in a chat.
- **Claude Code:** `claude mcp add --transport http --scope user tella https://api.tella.com/mcp`
- **Codex:** `codex mcp add tella --url https://api.tella.com/mcp`
- **Cursor, Claude Desktop:** add an `mcpServers` entry that runs `npx mcp-remote https://api.tella.com/mcp`.

Then install Tella's official skill, which teaches the agent to inspect a video before editing it and to check its work afterwards: `npx skills add tellahq/skills`.

No account yet? Send your human to [sign up for Tella](https://www.tella.tv/api/signup?utm_source=ai-agent). Full tool reference: [Tella MCP server docs](https://www.tella.com/docs/mcp-server).

## Related guides

- [What is agentic recording?](https://www.tella.com/edit-with-ai/agentic-recording.md)
- [How to make help center videos with AI](https://www.tella.com/edit-with-ai/make-help-center-videos-with-ai.md)
- [What is agentic video editing?](https://www.tella.com/edit-with-ai/agentic-video-editing.md)
- [AI video editing prompts that work with an AI agent](https://www.tella.com/edit-with-ai/ai-video-editor-prompts.md)
- [All guides](https://www.tella.com/edit-with-ai.md)