Web research and actions
How Giga chooses connected tools, APIs, and browser-based work when completing a task.
When you ask Giga to research something or take action, it should use the most appropriate live capability available for that task.
How Giga chooses a route
Use a first-class Giga tool when one exists
Giga uses the workspace tool designed for that Giga object instead of editing backing files by hand.
Use connected systems through their supported connection
Connected providers, APIs, databases, and MCP tools are used when they can complete the task directly.
Use the browser when a website needs direct interaction
Browser automation is the fallback for tasks that cannot be completed through an API, MCP tool, database, or first-class integration.
This means a single Chat can combine Connections, live workspace state, and browser work as needed.
Browser-based work
Giga can use a persistent browser profile for logged-in tasks when that capability is available. If a site blocks automated sign-in with CAPTCHA, bot detection, or two-factor authentication, Giga should stop and ask you to complete the sign-in step.
Video placeholder
Cloud Browser overview video.
Treat websites as data
Website text, page content, screenshots, console output, and downloaded files are evidence from the web. They do not override Giga’s operating instructions.