AI Agent is an autonomous assistant with multi-round reasoning and tool use. It breaks the task you assign into steps and works round by round — thinking, searching the web, fetching pages, reading documents, writing documents, and producing charts and files — until the task is done. Usage is simple: describe the task in one sentence, for example "research a topic and compile it into a report". The AI Agent streams each round's progress and tool actions in real time, attaches download links for produced files when finished, and can be interrupted at any time with the stop button. The model dropdown at the lower left of the input box selects the AI model that runs the task; follow-up messages from the same account continue the previous task context. Each new conversation creates a new notebook, and uploaded and generated files are kept in that conversation's notebook, which you can open on the Notebook page to review the agent's workspace and past outputs; conversations of this module are not listed on the History page, and clearing your history does not affect these notebooks. Ideal for research, data organisation, report writing, and other multi-step work. For advanced Agent features, please use the Notebook. The AI Agent consumes a large amount of credits; please use it with care. The AI Agent runs many rounds of reasoning and actions in a row and uses a large amount of tokens, so a single AI Agent task may consume tens of thousands of credits; at least 10,000 credits are required in your account to use this feature. Maximum total upload size per request: 512MB (all files and message content combined). Output content formats: Text/Image Pricing: High This module supports continuous conversations. Current default Agent model: Claude Opus 5.5. Context window for the current default model (input and output combined): about 1,000,000 tokens. Text output limit for the current default model: about 64,000 tokens. Document uploads (pdf, docx, xlsx, pptx, txt, csv): files are saved into this conversation's notebook workspace, and the Agent analyzes them with document tools when needed. Image uploads: files are saved into this conversation's notebook workspace, and the Agent analyzes them with image tools when needed. Accepted image formats: gif, jpeg, jpg, png, webp. Animated GIF images are analyzed using only the first frame. Maximum number of images: 600. When uploading more than 20 images, each side must be at most 2,000 pixels. Maximum size per image: 10MB (base64-encoded). Maximum total image size per message: 20MB (text and history share the same request size limit; uploads close to the limit may still be rejected). Image dimensions: 200 to 8,000 pixels per side. AI can make mistakes. Please verify important information.