Use the TokenHub API in the AI tools you already use.

Find setup guides for Claude Code, Cline, OpenClaw, and more. Enter your TokenHub API key, API base URL, and model ID as shown in each guide.

Find your app

What do you need to do?

Start with the issue you have and find the next step.

App setup guides

Find your app, then check where to configure it, which API it uses, and any conditions.

Choose a model that fits

Rule out incompatible models first, then compare real task results, speed, and cost.

  1. CompatibilityCheck API and features

    Verify the app’s required API and capabilities, such as Responses, Messages, or tool use.

  2. Task qualityTry your own work

    Compare output quality, speed, and context handling on typical tasks, not just one chat message.

  3. Ongoing costEstimate actual usage

    Factor in input, output, and request frequency before choosing a daily model.

Explore model capabilities and pricing

Make your first request

Follow your app’s guide for the exact fields. Then verify the setup with one simple request.

Where a request goes
Your AI toolAPI URL · Key · Model ID
TokenHubCheck the request log
Selected modelReply returns via TokenHub
  1. Check the app and API format

    Confirm where to configure it, whether it uses an OpenAI- or Anthropic-compatible API, and any feature limits.

  2. Get your key and model ID

    Create a TokenHub API key and copy the full model ID from the model catalog. Enter them in separate fields.

  3. Save and reload your settings

    Enter the API base, key, and model ID as shown in the guide. Restart the app or session if it reads environment variables or a config file.

  4. Send a simple test request

    Check both the app’s result and TokenHub request history. Test tool use and other required features separately.

Something failed? Start with the symptom

Send a simple request, then compare the app error, request time, and TokenHub request history.

Open the TokenHub console
What to check

No request appears

Check these in order, retrying after each change.

  1. Check that you are viewing the right account and time range.
  2. Confirm the app sent a test request using the custom provider.
  3. Check the API base and network; restart the app after config changes.
Find the app’s setup guide
No request appearsCheck the account, time range, and test

Check these in order, retrying after each change.

  1. Check that you are viewing the right account and time range.
  2. Confirm the app sent a test request using the custom provider.
  3. Check the API base and network; restart the app after config changes.
Find the app’s setup guide
Key or permission errorVerify the key, field, and account

Check these in order, retrying after each change.

  1. Make sure the API key is complete and active.
  2. Enter it in the API key field, not the model or URL field.
  3. Check account access; if no request appears, inspect the app settings first.
Find the app’s setup guide
Model or endpoint errorVerify the model ID, API, and URL

Check these in order, retrying after each change.

  1. Copy the full, available model ID from the model catalog.
  2. Select the OpenAI- or Anthropic-compatible API specified in the guide.
  3. Use the guide’s Base URL; do not add or remove /v1 or request paths by guesswork.
Find the app’s setup guide
Chat works, but tasks failTest the required capability

Check these in order, retrying after each change.

  1. Confirm the model supports the tool-use features your task needs.
  2. Check whether the app version, mode, and account enable that feature.
  3. Test chat and the target feature separately, comparing errors and requests.
Find the app’s setup guide
Request arrived, but no replyCheck limits, rate limits, timeouts, and service errors

Check these in order, retrying after each change.

  1. Find the request by test time and check its status, model ID, and error details.
  2. For a balance or rate limit error, check available credit and model limits, then retry later.
  3. For a timeout or service error, retry with a short prompt; if it persists, note the request time and error.
Find the app’s setup guide

Questions, answered

Can I use the same URL for OpenAI- and Anthropic-compatible APIs?

No. Choose the API format your app supports and follow its guide. Apps can handle /v1 and request paths differently, so do not copy a URL from another app.

What if my app does not show TokenHub models?

Some apps do not fetch models from custom providers. Copy the full model ID from the TokenHub catalog and add it manually. If there is no manual model field, check the app guide for limitations.

My app is not listed. How can I check compatibility?

Look for fields for a custom API base, API key, and model ID, then check the supported API format. An app that only allows sign-in to fixed providers cannot use another app’s setup steps.

Does TokenHub model usage include my app subscription?

No. TokenHub charges for model requests according to its prices and actual usage. The app may charge a separate subscription. Check both price pages before use.