OpenAI is switching off the models your automations run on.
We move your n8n, Make, Zapier, Python or Node workflows to the replacement model and test them on your own inputs, before and after, within 48 hours. Not fixed in 48 hours, full refund.
See pricingFree scan firstWhat shuts down, and when
| Shutdown | Affected | Replacement |
|---|---|---|
| 30 Nov 2026 | Agent Builder, /v1/prompts (reusable prompt objects) | Agents SDK; prompts in your own code |
| 11 Dec 2026 | gpt-5 family, o3 | gpt-6.1-sol and siblings |
| 14 Jan 2027 | gpt-4o, gpt-4o-mini, gpt-4.1, gpt-4.1-mini, including fine-tunes | gpt-6-luna |
Source: platform.openai.com/docs/deprecations, checked 14 Nov 2026. Anything still calling these stops working on the date.
How it works
- Free scan
Run the free scanner on your workflow export or code to see what is affected and how many workflows.
- Pick a tier
Choose by workflow count. After purchase, reply to the receipt with your export and a restricted OpenAI API key for testing.
- We migrate and test
Switched to the replacement model, plus a before/after run on up to 20 of your real inputs and the cost difference per 1,000 calls. Within 48 hours, or full refund.
Test tokens are billed to your own API key.
Pricing
1 workflow
- One workflow or script migrated
- Before/after test report
- 48 hours or full refund
Up to 5 workflows
- Up to 5 workflows or scripts migrated
- Before/after test report for each
- 48 hours or full refund
Up to 15 workflows
- Up to 15 workflows or scripts migrated
- Agent Builder rebuild
- Before/after test report for each
- 48 hours or full refund
Secure checkout opens on this page, processed by Gumroad.