Chat with LLM-controlled browser

Send a conversation to the LLM which uses browser tools to complete tasks. The LLM autonomously calls analyze_page, click, type, etc. until the task is done.

post/browsers/{browserId}/chat

Request

  • Base URL: https://oyabrowser.com
  • URL: https://oyabrowser.com/browsers/{browserId}/chat
  • Auth: HTTP bearer

Path parameters

browserIdstring required

Browser session ID (from GET /browsers)

Request body

dataobject

Task values the agent may read, referred to as {{name}} in the prompt; a value may be a file() to upload

secretsobject

Task values the agent never sees; typed through {{name}} placeholders

schemaobject

A JSON schema for the answer. The agent then finishes by returning data in this shape, answered as data

Response

Chat response

textstring

LLM's final text response

replayableboolean

Whether the run acted on a page and can be saved as a playbook (POST /api/browsers/{browserId}/playbooks)

failedboolean

Whether the agent reports it could not do the task (its reply starts with FAILED:, or it ran out of steps)

dataunknown

Changes

No recorded changes to this operation across all 1 revision of this API.