TKEN

Open WebUI provider setup

Connect Open WebUI to TKEN with an OpenAI-compatible base URL

Use TKEN as a third-party OpenAI-compatible API gateway for Open WebUI when you want one chat UI to test GPT, Claude, DeepSeek, Qwen, Kimi, MiniMax and other model families through a single endpoint.

Get API key Developer resources
Open WebUI connection TKEN values 
URL: https://www.tken.shop/v1
API key: store on the Open WebUI server
Model IDs: use IDs returned by /v1/models
OpenAI-compatible provider URL
Server-side API key handling
Preflight checks before shared use

Start with a terminal preflight

Before changing Open WebUI, verify that the same TKEN key can list models and complete one low-risk request. This separates key, account, and model-route problems from Open WebUI configuration problems.

cURL replace placeholders 
export TKEN_API_KEY="sk-your-tken-key"
export TKEN_BASE_URL="https://www.tken.shop/v1"

curl "$TKEN_BASE_URL/models" \
  -H "Authorization: Bearer $TKEN_API_KEY"

curl "$TKEN_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $TKEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "replace-with-an-available-model",
    "messages": [
      {"role": "user", "content": "Reply with one short sentence."}
    ]
  }'

Admin UI setup

1. Open the OpenAI connection settings

In Open WebUI, go to Admin Settings, open Connections, choose the OpenAI connection area, and add a new OpenAI-compatible provider connection.

2. Add the TKEN endpoint

Set the provider URL to https://www.tken.shop/v1 and use a limited TKEN API key created for Open WebUI traffic.

3. Restrict visible models when needed

If model discovery is incomplete or too broad, add only the model IDs you want users to see in Open WebUI's Model IDs filter.

4. Save, restart if required, and test one prompt

Select a known model ID, run one short prompt, and confirm that logs do not expose API keys, account details, or private prompts.

This page covers Open WebUI server-side provider setup. If you enable browser-side direct connections in Open WebUI, review the trust and key-exposure model separately before sharing the workspace.

Environment variable setup

For Docker, Kubernetes, or repeatable deployments, keep the TKEN base URL and key in server environment variables or deployment secrets. Do not commit real keys to public repos.

.env example server only 
ENABLE_OPENAI_API=True
OPENAI_API_BASE_URL=https://www.tken.shop/v1
OPENAI_API_KEY=sk-your-tken-key

# Optional after /v1/models returns your chosen route.
# DEFAULT_MODELS=replace-with-an-available-model
# TASK_MODEL_EXTERNAL=replace-with-a-fast-low-cost-model

# Keep passthrough disabled unless access control is reviewed.
# ENABLE_OPENAI_API_PASSTHROUGH=False

Production checklist

Endpoint and model checks

Confirm /v1/models and one non-streaming chat completion with the same key, base URL, and model ID that Open WebUI will use.

Key handling

Use a limited key for UI traffic. Avoid provider management keys, screenshots containing secrets, and browser-side key exposure.

Cost controls

Set default models deliberately, cap output tokens for shared users, and review live pricing before agentic or batch use.

Team rollout

Start with one user and one model route, then add more models only after response quality, latency, and usage behavior look sane.

Failure map

Most Open WebUI gateway issues come from mismatched keys, stale model IDs, or model discovery assumptions. Diagnose with the terminal preflight before changing multiple settings at once.

No models appear
Check /v1/models and model filters
401 error
Replace the Open WebUI server key
404 model error
Refresh model IDs from the gateway
Unexpected spend
Review pricing, defaults and token caps

Open WebUI and TKEN FAQ

What base URL should I enter?

Use https://www.tken.shop/v1 as the OpenAI-compatible provider URL in Open WebUI.

Can I use multiple OpenAI-compatible providers?

Open WebUI supports multiple connection patterns. Keep the key-to-URL pairing clear and test each route separately if you configure more than one provider.

Should I use a direct browser connection?

For shared workspaces, prefer server-side provider configuration unless you have reviewed how browser-side direct connections store and expose user keys.

Is TKEN an official Open WebUI or OpenAI service?

No. TKEN is an independent third-party OpenAI-compatible API gateway and is not officially affiliated with Open WebUI, OpenAI, or other upstream model providers.

Run Open WebUI through one verified gateway endpoint

Start with TKEN base URL https://www.tken.shop/v1, validate one model response, then use the developer hub for SDK smoke tests, cost guardrails, and related OpenAI-compatible setup guides.

Create account Developer hub