Chatbox Setup: Use Claude and GPT via a Custom API Provider

Chatbox (Chatbox AI) is a chat client for Windows, macOS, Linux, iOS, Android and the web. Besides its own paid Chatbox AI service, it lets you bring your own API key for many providers, including any service that speaks the OpenAI or Claude API format through a custom provider.

APIVAI works as such a custom provider. One APIVAI key gives you Claude and GPT models, billed per token below official list prices from a balance you top up, with no subscription. The same key works on your desktop and your phone, so you can set Chatbox up once on each device and see all usage in one Dashboard.

When you add a custom provider, Chatbox asks for an API Mode. Two modes fit APIVAI:

Pick OpenAI API Compatible if you want Claude and GPT in one provider; this is the simplest setup. Add a Claude API Compatible provider as well if you mainly use Claude: Chatbox only shows its thinking controls for a Claude model when it is reached through the Claude API style. The same key works in both. Chatbox also lists OpenAI Responses API Compatible; you don't need it for this guide.

Note the /v1 rule, which is different from many other tools: in Chatbox's Claude mode the API Host must include /v1, because Chatbox appends only /messages. Under the fields Chatbox shows a Preview of the final URL. It should read https://api.apivai.com/v1/chat/completions for the OpenAI mode and https://api.apivai.com/v1/messages for the Claude mode.

1. Sign up at apivai.com with your email.
2. Top up your balance from $10 by card, crypto, Alipay or WeChat Pay. There is no free trial; you pay only for what you use.
3. Open the Dashboard, create a key and copy it. You can set a budget limit for the key, which is useful for a phone app.

1. Open Settings → Model Provider. On mobile, tap the menu in the top-left corner to open the sidebar first.
2. Click Add and select Add Custom Provider.
3. Enter APIVAI as Name, choose OpenAI API Compatible as API Mode and confirm with Add.
4. Fill in the provider settings:

Check that the Preview line shows https://api.apivai.com/v1/chat/completions, with /v1 exactly once.

Under Models, click Fetch to load the models your key can use, or click New and enter one by hand:

Add one entry per model. Model IDs change over time, so copy them from Fetch, GET https://api.apivai.com/v1/models or the pricing page rather than from old guides.

Click Check to test the key and host, then pick the new model in the model selector of a conversation and send a message.

Repeat Step 2 with Name APIVAI Claude and API Mode Claude API Compatible. Use the same API Key and the API Host https://api.apivai.com/v1; leave API Path empty. The Preview should show https://api.apivai.com/v1/messages. Add only Claude models here, such as claude-sonnet-4-6 or claude-opus-5-5. The Claude API proxy page explains the Anthropic format.

Prices per 1M tokens as of 2026-10-07. All models are on the pricing page.

A good default for most people is Claude Sonnet 4.6. Switch to Claude Haiku 4.5 or GPT-6 Luna for quick, short tasks, and to Claude Opus 5.5 only when a question really needs it.

You pay only for tokens. Each message resends the conversation so far, so long threads cost more per turn. Thinking is on by default and thinking tokens are billed as output.

• Everyday chat: assume about 2,500 input tokens (your message plus recent history) and 700 output tokens per turn, 30 turns a day for 30 days (900 turns). On Claude Sonnet 4.6 that is $5.63 a month, against $16.20 at official prices. On Claude Haiku 4.5 the same month is $1.88, and on GPT-6 Luna $0.0808.
• Hard questions: 10 questions to Claude Opus 5.5 with about 8,000 input and 2,000 output tokens each cost $0.25 ($0.72 at official prices).

Start a new conversation when the topic changes, and limit how many earlier messages Chatbox sends if your version offers that in the conversation settings. The Dashboard lists every request with its tokens and cost. For a full comparison, see Claude API pricing compared.

Chat works fully through APIVAI, but a knowledge base needs an Embedding Model to index your files, and APIVAI does not provide embeddings. Configure a separate embedding provider for the knowledge base; the model that answers can still be Claude or GPT from APIVAI. Image generation is not available through APIVAI either.

• 401 or "invalid API key": the key was copied with a space or line break, has been deleted, or has reached its budget. Copy it again from the Dashboard and check your balance.
• 404: look at the Preview line. For the OpenAI mode it must be https://api.apivai.com/v1/chat/completions; a host without /v1 combined with an API Path, or /v1 twice, gives 404. For the Claude mode it must be https://api.apivai.com/v1/messages; https://api.apivai.com without /v1 gives 404 here.
• Model not found: the Model ID must match exactly, for example claude-sonnet-4-6. GPT models fail in a Claude API Compatible provider; add them to the OpenAI one.
• Empty or cut-off replies: thinking tokens count towards the output limit. Raise Max Output Tokens in the model entry or in the conversation settings to 4096 or more.
• Connection errors or broken streaming: turn on Improve Network Compatibility in the provider settings and try again.
• 429: the default limit is 60 requests per minute per key; wait a moment or ask support for a higher limit.

The docs list all endpoints and error codes.

Yes, in both modes: https://api.apivai.com/v1. Chatbox adds /chat/completions (from API Path) or /messages after it. Check the Preview line.

No. A custom provider uses your own key, and APIVAI bills per token from a prepaid balance. Chatbox's own paid plans are a separate service.

Yes. Add the same custom provider in the Chatbox mobile app. One key works on all devices; give it a budget limit if you want a cap.

Yes. One APIVAI key works for every model on the pricing page, in both the OpenAI and the Claude format. It also works in Cherry Studio and other OpenAI-compatible tools.

No. Request and response content is not logged; only usage data such as model, token counts and cost is kept for billing.

Create an account, top up from $10 and paste your key into Chatbox.