Channel
Talk to your AI agent in the browser
Chat streams both the answer and what the agent is doing to produce it, so a run that takes thirty seconds is legible rather than a spinner.
What you get
Web chat capabilities, under your rules
- ✓ Streams the answer and the agent’s progress as it works
- ✓ The same runtime, tool registry, policy engine and approval gate as every other channel
- ✓ Runs as the signed-in person, under their role’s permissions
- ✓ Every run recorded as a trace you can read afterwards
What you need to connect it
- A Caterfli account in the workspace
- At least one agent
Human approval
Destructive work waits for a person
approval
waiting
delete_customers
CRITICAL risk
Destructive
- Tool
- delete_customers
- Operation
- DELETE
- Runs against
- Production PostgreSQL
- Agent
- Support Agent
Exactly these values
{
"filters": {
"id": 100
}
}
Your decision applies to these values only. If anything about the call changes, it stops for approval again.
A destructive request stops for a person. The approval covers this call with these arguments, once.
Connect Web chat to your first agent
Connect one system and read the agent Caterfli writes from it. There is a demo connection too, if you would rather look before you wire anything up.