frontend-api-demos

To create a conversation, send a fetch request to the /api/demos/conversations endpoint, optionally including an agent_id. This initiates a new dialogue session for the chosen assistant, enabling subsequent message sending within your UI.

Yes, you can upload a PDF by sending a fetch request to the /api/demos/upload-pdf endpoint. This uploads the file and attaches its content to a conversation, allowing the assistant to process the document within the demos service.

To send messages to an ongoing conversation, use the fetch API to target the /api/demos/conversations/:id/messages endpoint. This continues the dialogue by transmitting user input to the specific session identified by its ID.

No external dependencies are required to integrate the demos API. The solution relies solely on standard browser fetch capabilities, allowing frontend developers to test API interactions and handle loading states without installing additional libraries.

The best way to handle errors and loading states is to manage them according to the documented API responses. Use standard fetch error handling in your frontend to gracefully manage request failures and UI loading transitions during API interactions.