Agents
Describe what you need done and let autonomous AI agents handle the rest. They research, plan, execute multi-step tasks, and deliver results.
Tell it what you want. Watch it work.
How Agents Work
1. Describe What You Need Done
Write a description of your task in plain language. Attach files for context — PDFs, images, documents, spreadsheets. The more detail you provide, the better the results.
Advanced Configuration
- Choose your models — select which AI models the agent can use (ChatGPT, Claude, Gemini, Grok) and pick the planner and coordinator models
- Enable your connectors — give the agent access to your Google Calendar, Gmail, Drive, Slack, and other connected apps
- Attach files — upload up to 10 files (images, PDFs, docs) for the agent to reference
Describe your task
Configure
Launch
2. The Agent Plans & Executes
Once launched, the agent works autonomously. A Planner model breaks your request into a structured plan of tasks. Then a Coordinator orchestrates execution — assigning each task to the best AI model and tools for the job.
Planner
Breaks task into sub-tasks with dependencies
Coordinator
Assigns models, tools, and executes each task
Task 1
Research
Task 2
Analyse
Task 3
Write report
Multi-model orchestration — the agent can use different AI models for different tasks. Claude for analysis, ChatGPT for writing, Gemini for research — whatever works best.
3. Watch Progress in Real-Time
Follow along as the agent works. See the planner thinking, tasks executing, tool calls happening, and results streaming in — all live.
Live Streaming
Watch thinking, tool calls, and outputs as they happen
Cost Tracking
Real-time token usage and cost breakdown per task
Tool Visibility
See every web search, API call, and file operation
4. Interact & Steer
You're always in control. Pause, resume, or cancel at any time. Send messages to redirect the agent mid-execution. Respond to clarification requests when the agent needs your input.
- •Clarifications — the agent asks you when it needs more context or faces ambiguity
- •Approvals — optionally require your sign-off before the agent takes critical actions
- •Follow-ups — send messages to completed agents to continue or refine their work
5. Get Your Results
When the agent finishes, all outputs are saved to your Notebook — documents, reports, files, and deliverables. Get notified via app or email when it's done.
Competitor Analysis Report
5 tasks · $0.12 · 2 min ago
Q1 Revenue Summary
3 tasks · $0.08 · 1 hour ago
Blog Post Draft: AI Trends 2026
4 tasks · $0.15 · Yesterday
Why Agents
- Autonomous execution — describe what you need and let the agent figure out how to get it done
- Multi-model orchestration — agents use different AI models for different tasks, picking the best tool for each job
- Connected to your apps — agents can use your connectors to search Gmail, check Calendar, read Drive files, and more
- Full transparency — watch every step live, see tool calls, thinking, and costs in real-time
- Always in control — pause, resume, cancel, or redirect at any point during execution
What Can Agents Do?
Deep Research & Reports
Research any topic across the web and compile a structured report with sources and analysis.
Competitive Analysis
Analyse competitors, compare features, pricing, and market positioning — delivered as a ready-to-share document.
Content Creation
Research a topic, outline, draft, and refine — blog posts, newsletters, presentations, and more.
Data Analysis
Upload spreadsheets or connect to your apps, then let agents analyse, visualise, and summarise your data.
