davsclaus opened a new issue, #3361: URL: https://github.com/apache/camel-kamelets/issues/3361
A new action Kamelet, `openai-chat-action`, on the Camel `openai` component (Stable since 4.17): it sends the message body to a chat model and gives back the answer. - **Any OpenAI-compatible server:** `baseUrl` lets it call OpenAI or a local server such as Ollama, vLLM or LM Studio. `apiKey` is a credential. - **Options:** `model`, `systemMessage`, `temperature`, `maxTokens`, and `stripThinking` for reasoning models. - **Structured output:** with `jsonSchema`, the answer is JSON text that follows the schema, for classification and extraction. - **Input:** the body is the user message as text. **Output:** the answer as text, or the JSON text. - **Security:** the message is untrusted. The Kamelet does not let the model call tools (no `tags`), so message text cannot make the model act. It also removes the component's own headers from the incoming message, so a message cannot change the model or the request. - **Citrus tests:** an HTTP stub of `/v1/chat/completions`, with no real model and no API key. The tests check a plain answer, the request the stub gets (model, system message, user message), a `jsonSchema` answer, and an error from the server. It sits next to `aws-bedrock-text-sink` and `google-vertexai-sink`. _Claude Code on behalf of Claus Ibsen_ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
