Keep LiteLLM's router, change what it bills to.
LiteLLM is great for internal model access. Add ProxyLLM as the upstream and the OpenAI traffic behind it bills to your ChatGPT subscription via Codex Hosted, with key custody and hard spend caps.
$129/month SaaS. Bring your own model keys. No inference markup.
Three steps to connect.
Create an upstream key
Generate a ProxyLLM key for your LiteLLM proxy. Keep the cap separate from direct app keys so router traffic is easy to audit.
Add ProxyLLM as a model backend
Use https://api.proxyllm.ai/v1 as the api_base for OpenAI-compatible traffic and pass your ProxyLLM key as the upstream credential.
Add budgets above the router
LiteLLM keeps its team router. ProxyLLM holds the upstream credential, logs every request, and serves OpenAI models from your ChatGPT plan with a hard cap.
Add ProxyLLM as an upstream.
Point a LiteLLM model at the OpenAI-compatible ProxyLLM endpoint.
model_list:
- model_name: proxyllm-gpt-4o-mini
litellm_params:
model: openai/gpt-4o-mini
api_base: https://api.proxyllm.ai/v1
api_key: os.environ/PROXYLLM_API_KEYRun 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.
A second line of defense for model spend.
Use LiteLLM for internal teams. ProxyLLM holds the upstream credentials, caps them, and bills OpenAI traffic to your flat ChatGPT subscription.