Bring AI models
into your product.

OpenAI, DeepSeek, and GLM through one API.
Separate keys and usage tracking for bots, apps, and agents.

CONNECTION01 / 03

A model for your task

Your app
Familiar tools

Python · JavaScript · cURL

A graphite sphere with silver connections, illustrating one point of access to AI models.
ONE ENDPOINT

Different models.
One API.

api.wikivibe.dev/v1
Spending by keyExample
$24.80 / $100

Every project
gets its own key

Manage access and quotas in the console.

Browse models OpenAI-compatible API

Choose a model.
Compare the cost.

Compare input and output prices side by side.
Choose a model to get a request example
with its ID already filled in.

Give each project
its own key and budget.

Give your support bot and internal agent separate limits. Manage access and see where your balance goes.

A key for each project

Grant access to the model group you need.

Usage quota

Set a budget for each key.

Cost per request

Review input tokens, output tokens, and charges.

Spending by projectConsole preview

Budget used

$24.80/ $100.00
24.8% used$75.20 remaining
SSupport bot$16.42
AInternal agent$6.18
DDevelopment$2.20
Illustrative projects and amounts.

Your first request.
Keep your current SDK.

  1. Create a key in the console

    Choose the model group for your project.

  2. Set the WikiVibe endpoint

    Set the base URL and API key in your client.

  3. Send a request

    The example already includes the model ID.
    Read the docs

Choose a modelChange model

What else
should you know?

Do I need to rewrite my app?

If your client supports the OpenAI API format, set https://api.wikivibe.dev/v1 as the endpoint and use a WikiVibe key and an available model ID. Supported methods and parameters depend on the model. Check the documentation.

Can I use one key for every model?

Create a key for the model group you need. When switching providers, check the key's group and permissions. The API endpoint stays the same.

How are requests priced?

Costs depend on token usage and model rates. Input and output have separate rates. Cache usage, context size, paid tools, and processing mode may also affect the cost. Prices come from the API. See actual charges in the console.

Now build with it.

Choose a model, create a key, and send your first request.

Go to the console