Zaps that call AI, on a flat AI bill.
Zapier runs unattended across thousands of tasks, which is exactly where per-token billing stings. A Webhooks by Zapier step points at ProxyLLM, and its OpenAI calls run through Codex Hosted on your ChatGPT subscription, behind a scoped key with a hard cap.
Free for 7 days, then $129/month. Bring your own model keys. No inference markup.
Three steps to connect.
Create an automation key
Generate a ProxyLLM key for Zapier and set a monthly cap. A Zap that loops or retries should have its own spending boundary.
Add a Webhooks by Zapier step
Zapier's native OpenAI app has no base-URL field, so use the Webhooks by Zapier action (a paid Zapier feature). Pick Custom Request or POST, and send to https://api.proxyllm.ai/v1/chat/completions.
Set the header and body
Add an Authorization header of Bearer plus your ProxyLLM key, send the chat-completions JSON, and each Zap request is attributed to that key so you can see automation spend on its own.
POST to the chat completions endpoint.
Zapier's native OpenAI app can't set a base URL, so a Webhooks by Zapier Custom Request carries the ProxyLLM key and body.
POST https://api.proxyllm.ai/v1/chat/completions
Authorization: Bearer pk_live_...
Content-Type: application/json
{
"model": "gpt-4o-mini",
"messages": [{ "role": "user", "content": "Summarize this email." }]
}Run your AI workloads on your ChatGPT subscription.
ProxyLLM runs OpenAI's Codex for you, signed in with your own ChatGPT account. Your apps call one OpenAI-compatible endpoint and the work bills to your flat plan instead of per-token API pricing.
Automate without open-ended spend.
Classification, drafting, and extraction steps bill to your ChatGPT subscription through Codex Hosted. When a plan limit hits, your own OpenAI key picks up until it resets.