Create an SOP with a video of every step

Describe a process and Tella Auto Capture films it in your web app. Claude or ChatGPT then writes the SOP from the video, with a screenshot for each step.

Auto Capture filming "record a customer's payment" in Tallyfern, a made-up invoicing app we test with, with one chapter per SOP step.

Try it in Tella

Paste a prompt into the Tella Agent in the editor, or into Claude or ChatGPT connected to Tella, then ask for changes.

Start here
Write an SOP for issuing a partial refund in https://app.example.com. Use Tella Auto Capture to film the process, signed in with the "Ops sandbox" credential: open an order, click Refund, enter $20, pick the reason "Damaged item" and stop on the confirmation screen without submitting. When it's delivered, write the SOP as numbered steps from the take's chapters, add a frame from the video for each step, and put the take in a Tella video named "SOP: Issue a partial refund".
  • From steps you already have

    Here is our checklist for setting up a new client in https://admin.example.com: [paste the steps]. Use Tella Auto Capture to follow it step by step with the "Admin demo" credential and tell me any step that doesn't match the app. Then rewrite the checklist as an SOP from what it filmed, with the Tella video linked at the top.

  • Ops team library

    For each of these tasks in https://app.example.com, film one Tella Auto Capture take with the "Ops sandbox" credential: add a user, change a user's role, export last month's invoices. Work through them one at a time. Write a one-page SOP for each with numbered steps and a frame per step, and name each Tella video after its SOP.

How can I create an SOP with AI, with a video and a screenshot of every step?

Ask Claude or ChatGPT, connected to Tella's MCP server, to film the process with Auto Capture and write the SOP from the take. Auto Capture runs the steps in its own browser, signed in with a saved login, and delivers a video with a chapter for each section. The AI writes the numbered steps from those chapters, takes a frame of each step from the video and links the Tella video at the top.

How to do it

  1. Save a login for the app in Settings → Credentials, ideally a sandbox account.
  2. Describe the process in your prompt, or paste the steps you already have.
  3. Ask the AI to film it with Auto Capture and write the SOP from the take's chapters.
  4. Read the handoff note for anything done off camera, then share the SOP with the video.

Good to know

  • Beta, on paid plans. Each take uses your workspace's credits, the same balance the Tella Agent spends.
  • What a take saves is saved for real, on every practice run too. Film processes that change data on a sandbox account, or stop before the final submit.
  • One take films one process. For a set of SOPs, ask the AI to work through them one at a time.
  • Auto Capture films web pages on the public internet, not desktop or mobile apps, localhost or intranet hosts.

Frequently asked questions

Yes. An AI app connected to Tella reads the take's chapters and handoff note and looks at frames from the video, so each written step matches what was filmed.

Ask for a new take of the same process and have the AI rewrite the steps that changed. A revised take starts from what the earlier one learned, so it costs less and finishes sooner than starting over.

Auto Capture follows what you wrote where it can. The handoff note says where it departed from your request, which usually points to a step that's missing or out of date.

Get your next demo filmed for you

Get started