Frameworks & Platforms
Portkey Gateway
Connect Portkey Gateway to TokenLab
Overview
Configure TokenLab as a custom OpenAI-compatible host in Portkey. Use Responses only when your Portkey version supports a custom Responses endpoint.
Provider Values
| Field | Value |
|---|---|
| Provider | openai |
| Custom Host | https://api.tokenlab.sh/v1 |
| API key | sk-your-tokenlab-key |
| Model | A current TokenLab model ID |
Chat Completions
Use TokenLab for OpenAI-compatible chat requests:
The example requires a running self-hosted Portkey Gateway; replace the gateway URL with your deployment URL. For Portkey Cloud, also send x-portkey-api-key with your Portkey API key. See the Portkey configuration guide.
curl https://your-portkey-gateway.example/v1/chat/completions \
-H "Content-Type: application/json" \
-H "x-portkey-provider: openai" \
-H "x-portkey-custom-host: https://api.tokenlab.sh/v1" \
-H "Authorization: Bearer sk-your-tokenlab-key" \
-d '{
"model": "gpt-5.6-luna",
"messages": [{"role": "user", "content": "Say hello from TokenLab."}]
}'Responses
If your Portkey version supports a custom Responses URL, choose a current Responses-capable model from the Models page and use /v1/responses.
Other API Formats
Anthropic Messages and Gemini can be used through Portkey only if your Portkey version supports custom URLs for those formats.