Give every Dify app its own model budget.
Dify apps start cheap and get expensive as workflows grow. ProxyLLM gives each app a scoped key with a hard cap, and OpenAI-model calls bill to your flat ChatGPT subscription through Codex Hosted.
$129/month SaaS. Bring your own model keys. No inference markup.
Three steps to connect.
Create a workspace key
Generate a ProxyLLM key for Dify and set a cap that matches the app or team using it.
Add an OpenAI-compatible provider
In Dify, add an OpenAI-compatible model provider with API Base set to https://api.proxyllm.ai/v1 and the ProxyLLM key as the API key.
Pick models for each app
Use plain OpenAI names or prefixed models like anthropic/ and google/. OpenAI models run through Codex Hosted on your ChatGPT plan; prefixed models go through your OpenRouter key.
Add one OpenAI-compatible provider.
Dify keeps its app builder flow. ProxyLLM handles keys, logging, and where each call gets billed.
Provider: OpenAI-compatible
API Base: https://api.proxyllm.ai/v1
API Key: pk_live_...
Model: gpt-4o-miniRun 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.
Ship Dify apps without blind spend.
Cap each workspace key, follow spend per app in the request log, and bill OpenAI traffic to your ChatGPT subscription instead of a metered key.