How to use Chocolate Factory.
Here is exactly how an AI workforce comes to life on Chocolate Factory: six steps, from creating your organization to governing every run in production, with no code required to get there.
Six steps, one continuous path.
No throwaway prototypes and no guesswork. Each step builds on the last: set up your workspace, compose an agent, connect your systems, orchestrate, execute, and govern.
Create your organization and project
Spin up a shared workspace for your team, invite owners and members, and provision the API keys that govern who can access what.
- Shared team workspaces
- Owner & member roles
- API keys to govern access
Compose an agent, no code required
Give an agent its role, instructions, and model, then ground it in your knowledge, all through a no-code setup, no engineering required.
- Its own role, instructions & model
- No-code / low-code setup
- Grounded in your knowledge
Equip it with your tools and data
Plug agents into MCP servers, databases, and knowledge bases so they can read your data and act on the systems you already run.
- MCP servers & databases
- Knowledge bases
- Reusable workflows
Chain agents into workflows
Compose multiple agents into a governed, multi-step workflow, with human approvals and guardrails on the steps that carry real risk.
- Multi-step agent workflows
- Approvals on risky steps
- Guardrails you define
Put your agents to work
Run agents from an in-app chat, or call them programmatically through the /chat and /execute APIs, on web, Teams, or WhatsApp.
- Chat interface, in-app
- /chat & /execute APIs
- Web, Teams & WhatsApp
Monitor and govern every run
Trace every execution in real time, track tokens and cost, and keep an immutable audit trail behind everything your agents do.
- Real-time execution traces
- Token & cost tracking
- Immutable audit trail
Want to see the whole platform?
Book a demo and we'll walk you through it.