Skip to content
Weik.ioCloud
Sign inCreate a workspace

Weik.io Cloud

AI models

AI models in Weik.io Cloud answer through an OpenAI-compatible API with keys per workspace and usage per key, and your Camel integrations call them with the HTTP or LangChain4j components.

What you get

With the AI models service you get:

  • An OpenAI-compatible endpoint at https://api.weikio.cloud/v1
  • API keys for each workspace, each with a name and an optional monthly limit
  • The usage of each key by day and by model in the console
  • An example Camel integration that extracts the fields of an incoming invoice

Sizes and prices

You pay for the tokens each key uses, with the usage of every key listed in the console under Billing.

AI models, prices per million tokens
ModelProviderPrice
Claude Sonnet 4.5, inputAnthropic, served through Google Vertex AI€3.00 per million tokens
Claude Sonnet 4.5, outputAnthropic, served through Google Vertex AI€15.00 per million tokens

Prices are in euros and exclude VAT.

How it connects to Weik.io

An integration calls the endpoint with Camel's HTTP component or with the LangChain4j components, and the key stays a secret in the workspace, such as ai-key in project finance.

The integration below receives the text of an incoming invoice from a channel and asks the model for the invoice's fields. It publishes the fields into the channel invoice-fields, where the integration that writes to your ERP receives them.

The example is a Camel route in YAML, and a Weik.io you install yourself runs it too.

finance/invoice-fields/invoice-fields.camel.yaml
- route:
    from:
      uri: "kamelet:weikio-event-source?channel=invoice-text"
      steps:
        - setVariable:
            name: invoice
            simple: "${body}"
        - setBody:
            simple: >-
              {"model": "claude-sonnet-4-5", "response_format": {"type": "json_object"},
              "messages": [{"role": "user", "content": "Return the invoice number,
              the date, the supplier and the total as JSON. ${variable.invoice}"}]}
        - setHeader: { name: Authorization, simple: "Bearer {{ai-key}}" }
        - setHeader: { name: Content-Type, constant: application/json }
        - to: "https://api.weikio.cloud/v1/chat/completions"
        - setBody: { jsonpath: "$.choices[0].message.content" }
        - to: "kamelet:weikio-event-sink?channel=invoice-fields"

The key is the secret ai-key, set with weikio secret set finance/ai-key --env production. Variables and secrets in the docs shows how an integration uses them.

Where it runs

API endpoint
Hetzner's Helsinki data centre park in Finland
Model provider
Google Vertex AI, listed in the subprocessors

Questions

Which client libraries work?

Any OpenAI-compatible client, with https://api.weikio.cloud/v1 as its base address, and Camel's LangChain4j components.

How is usage billed?

Per million tokens, input and output separately, with the usage of each key listed in the console.

Can I stop a key from spending too much?

Yes, give the key a monthly limit in the console. A call over the limit gets HTTP 429 until the next month or until you raise the limit.

Create an API key

The console creates the key and shows the endpoint and the example integration that extracts the fields of an invoice.