TKEN

LiteLLM gateway config

Route LiteLLM traffic through the TKEN gateway

Use TKEN as an upstream OpenAI-compatible endpoint when you want LiteLLM to sit between your apps and multiple model families while keeping one local proxy URL for client code.

Get API key Gateway overview
litellm_config.yaml OpenAI-compatible upstream 
model_list:
  - model_name: tken-minimax
    litellm_params:
      model: openai/MiniMax-M2.7
      api_base: https://www.tken.shop/v1
      api_key: os.environ/TKEN_API_KEY
One upstream base URL for TKEN
Keep app clients pointed at LiteLLM
Use smoke tests before production routing

Proxy start pattern

After the upstream model is configured, start LiteLLM with your config file and send local clients to the LiteLLM proxy. Keep production keys out of checked-in files and CI logs.

Local test replace secrets 
export TKEN_API_KEY="replace-with-your-key"
litellm --config litellm_config.yaml --port 4000

curl http://localhost:4000/v1/chat/completions \
  -H "Authorization: Bearer local-test-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tken-minimax",
    "messages": [{"role": "user", "content": "Reply in JSON."}]
  }'

Production-readiness checks

1. Confirm provider syntax in your LiteLLM version

LiteLLM changes quickly. Verify the current custom OpenAI-compatible provider syntax before copying a config into a production router.

2. Add a health-check request

Test /v1/models or one chat completion through LiteLLM and TKEN before sending user traffic through the route.

3. Keep logs safe

Redact authorization headers, prompts, customer identifiers and request bodies before saving logs or sharing debugging output.

Disclosure: TKEN is an independent third-party API gateway. It is not an official LiteLLM, OpenAI, Anthropic, Google, DeepSeek, Qwen, Kimi, MiniMax or GLM endpoint.

Put TKEN behind your LiteLLM proxy

Start with one route, confirm model behavior, then add fallback or cost controls only after the base request works.

Create account Pricing guide