Welcome to Glow

Glow is an agentic automation platform with built-in human governance. Combine AI agents, deterministic business logic, and human approvals on one live canvas. AI agents reason and call permitted tools, while deterministic rules run the same way every time.

Evaluating Glow for your team? Start with the platform review path for architecture, native integrations, access controls, and operations.

Choose Your Starting Point

Learn the canvas in about five minutes. Place a Scheduler and AI Prompt, run them manually, and read the result. Leave the automation in Draft. No external app account or code required.

Build your first automation

Prefer to describe the process? Build with Workflow Assistant in the app or use worked prompts.
Want to build from Claude, ChatGPT, Gemini, or your IDE? Connect your AI tool in 30 seconds and watch your Glow canvas update live as you chat.
Moving from n8n, Make, Zapier, or Power Automate? Follow Migrating to Glow to import or rebuild automations without missed events.
Need an account or a teammate? Set up Account & Authentication or try Working Together.
Delivering to a client? Start with Working With Clients and Your First Client Automation.

The Glow Difference

Build together in real time

Teammates edit and connect Steps on the same automation at the same time. See each other’s changes on one shared canvas.

Choose the tools an agent may use and which calls need approval. Use exact rules for routing, with Human Review to gate downstream actions that need sign-off.

Govern AI agents with rules and human review

Consultants and MSPs can build, deploy and monitor across governed client workspaces without mixing automations, accounts or run history.

Operate every client in a separate workspace

Describe the process in the in-app Workflow Assistant, or connect Claude, ChatGPT, Gemini, Cursor, or VS Code over MCP while your canvas updates live.

Build with Workflow Assistant or your own AI app

For the full product model, continue to What Is Glow?.

Learn by Building

After the quickstart, follow the tutorials in order:

Tutorial 2: Build It Yourself: map data between Steps, using Slack or the no-connected-account AI Prompt path.
Tutorial 3: Branching: connect Slack and Google Sheets, then test each route in Draft. Manual tests perform real actions, so use test destinations.
Tutorial 4: Your First AI Agent: build a live web assistant with conversation memory and hosted chat.

Go Further

Developer Hub & MCP: connect ChatGPT, Claude, Gemini, or an AI editor over MCP, write Python transforms, and test with execution tools.
How Glow Integrations Work: see how our integration framework connects app Steps, triggers, and agent tools.
Use Cases & Examples and Templates and Reuse: adapt worked examples or fork reusable templates.
Migrating to Glow, Operating Automations, and Troubleshooting: import workflows, switch to Live safely, and diagnose issues.

Verify it works

Run: Sign in at app.getglow.ai, open Automations → Create new → Workflow, and follow Tutorial 1 to run a Scheduler and AI Prompt in Draft
You should see: The AI Prompt Step's Executions tab records your first run and shows the generated note under result.

What’s Next?

👉 Build Your First Automation →