Tools in Chat
Web search, image generation, data sources, company search, and agent mode available to the model in chat
A model does not only answer from its training. The Tools menu below the input field gives it extra abilities: current information from the web, images, your own documents, connected apps, and a sandbox for running code. You switch a tool on, and the model reaches for it when it needs to.
Which tools you see depends on your organization and the selected model. Not every model supports every tool: image input, web search, and image generation are tied to the model's capabilities.
Web Search
Turn on Web search so the model researches live on the internet for current or factual questions instead of answering from its training cutoff. The sources it used appear as references in the answer. Ideal for news, prices, releases, or anything that may have changed since the model was trained.
Image Generation
Ask the model to create an image, and with Image generation enabled it produces one right in the chat. You can then refine a generated image further in the conversation. This tool requires an image-capable model in your workspace.
Data Sources
Under Data sources, the model works with content you provide yourself:
- Uploaded files: Attach PDFs, Word documents, images, or text files to a message. The model reads the content and refers to it directly.
- Document folders: Select whole folders that the model searches semantically, instead of attaching individual files.
- Project and app files: Files stored in a project or custom app are automatically available to every conversation there.
Large documents are indexed on upload and become searchable after a short moment.
Company Search
With connected integrations, the model searches the systems where your knowledge already lives and reads individual entries in detail. Connect an app once under Integrations; after that you attach content via Attach from integrations or let the model search on its own.
Full-text search across pages and read individual pages in detail.
Find issues by search and read a ticket in full.
Search files and folders and read file contents.
Search messages and channels by content.
Search emails and calendar entries (Gmail and Google Calendar where enabled).
Search pages and databases by content.
Agent Mode
In Agent mode the model gets a secure sandbox where it works through several steps on its own: it runs code, creates and edits files, and presents results for download. This solves tasks that go beyond a plain text answer, such as analyzing data from a spreadsheet, computing a chart, or producing a finished document.
Actions against external systems, for example via connected integrations, may be presented for approval before agent mode runs them. Review these steps before you confirm.
Artifacts and Code
Even without a dedicated tool in the menu, the model structures its outputs:
- Code blocks: Syntax-highlighted code with a copy button.
- Artifacts: Interactive outputs like charts, data visualizations, or rendered HTML appear in a dedicated panel beside the chat.
More Capabilities
Depending on what your organization has enabled, further tools are available. These include a personal memory where the model keeps notes about you across conversations, and scheduling tasks, which lets you receive recurring or timed results by email.