Chatbox: connect your own Zukko API key

Chatbox is a chat client for desktop and mobile where you can plug in your own API key instead of the built-in services. Zukko is added as an "OpenAI API Compatible" provider, after which Claude, GPT and GLM models become available in Chatbox with pay-per-token billing. Below is the step-by-step setup based on the official Chatbox docs, the details of the API Host and API Path fields, and fixes for common errors.

How to add your own API key to Chatbox

The Chatbox docs give a single set of steps, without separate desktop and mobile versions:

1. Open Chatbox and click the settings button in the sidebar.
2. Choose Model Provider.
3. At the bottom of the provider list, click Add.
4. Enter a name (for example, Zukko) and pick the API type OpenAI API Compatible.
5. Fill in API Key and API Host (values below). Leave API Path empty.
6. Add at least one model: type its ID manually, or pull the list with the Fetch button if your version has it.
7. Save and click Check — Chatbox should report a successful connection.
8. Go back to the main screen, start a new conversation, and select a Zukko model.

What to enter

Why you should leave API Path alone

Chatbox builds the final URL from two parts: API Host plus API Path. If the path is empty, the app uses /chat/completions, and if the host has no /v1, it adds one. So both host variants — https://api.zukko.pro/v1 and https://api.zukko.pro — lead to the correct https://api.zukko.pro/v1/chat/completions.

The error happens when /v1 is in both the host and the path: for example, host https://api.zukko.pro/v1 with path /v1/chat/completions produces /v1/v1/... and a 404. If you did fill in API Path, make sure /v1 appears exactly once in the final URL.

Models and their capabilities

Models in Chatbox are added manually by ID, exactly as they are named in the API. For each model you can tick capabilities (such as image input or tool use). Per the Chatbox docs, a model with nothing ticked is treated as text-only. Only tick capabilities when you're sure the model supports them, otherwise requests with images or tools will fail.

For the current list of IDs, call GET https://api.zukko.pro/v1/models (see the models documentation↗) or open the "Models" page in the dashboard.

An option for Claude: the Claude API Compatible provider

The type list also includes Claude API Compatible. Zukko accepts the Anthropic format at POST https://api.zukko.pro/v1/messages, and for this type Chatbox appends only /messages to the host. So the API Host here also needs the /v1 address: https://api.zukko.pro/v1.

You don't have to do this: the gateway converts formats, and Claude models work through the regular OpenAI-compatible provider. The second type is useful if you want to keep Claude models in a separate group. The difference between the formats is explained in the format selection docs↗.

Which model to choose

A sample calculation for one message: 2,000 tokens of question plus history and 800 tokens of answer, 2,800 in total:

• claude-sonnet-5: 2,800 × $0.3 / 1,000,000 = $0.00084;
• gpt-5.6-luna: 2,800 × $0.1 / 1,000,000 = $0.00028;
• glm-5.3-flash: 2,800 × $0.04 / 1,000,000 = $0.000112.

Check answer quality on your own tasks: add two or three models to Chatbox, ask each the same 10–20 typical questions, and compare the answers and spending in the dashboard's "Usage" section.

Desktop and phone

Provider settings are stored in the app on each device, so you'll need to add Zukko separately on your computer and your phone. It's convenient to create a separate key per device in the dashboard: if a phone gets lost, you delete one key without touching the others. Each key can have its own daily and monthly limit.

Common errors

401 or "Invalid API key"

The key was copied partially, with a space, or was deleted in the dashboard. Copy it again — it starts with fetch_.

404

Almost always the address: /v1 appears twice (in host and path), or there's a typo in the host. Clear API Path and set the host to https://api.zukko.pro/v1.

Unknown model error

The model ID doesn't match the API name: for example, Claude Sonnet 5 instead of claude-sonnet-5. Copy the ID from the GET /v1/models response.

400 on Claude models

claude-sonnet-5 and claude-opus-5-5 may reject a request with a 400 error if temperature is anything other than 1 or top_p is below 0.99. If you changed temperature or top_p in the conversation settings, restore the defaults for Claude.

402 or 403

402 means the balance ran out; top it up in the dashboard. 403 model_not_allowed_for_key means the model isn't on this key's allowed list; add it in the key settings.

Check passes but no answer arrives

See whether the model has capabilities ticked that it doesn't have, and whether the key limit has been reached. If the model provider is temporarily unavailable, the API returns 503 — pick another model or try later.

FAQ

Does this work on iOS and Android?

Yes, Chatbox has mobile versions, and the docs give one general set of steps: settings → Model Provider → Add → OpenAI API Compatible. If the items are named differently in your version, look for the model provider section.

Can I add several providers?

Yes, Chatbox can hold several providers, and you choose the model per conversation. For example, Zukko with one key for work and another for personal use.

How do I see spending?

In the Zukko dashboard under "Usage". Chatbox itself doesn't show your Zukko balance.

Where do I get a key?

Create one in the Zukko dashboard↗ and paste it into the API Key field. If APIs are new to you, start with the quick start guide↗.