Exact triggers
Runs start on the events you describe, and nothing else.
if: priority == "P1"
Each project gets its own agents. Every action they take passes the same deterministic checks.
Describe the job. Hinzel drafts and tests it. Nothing goes live until a person approves.
In Slack or web chat, the way you would ask a colleague.
It writes the rule, runs the tests and replays last week's events.
A project admin reviews the change. Then it runs the same way every time.
Every call an agent makes goes through one gateway. Rules you can read allow it, send it to a person, or block it.
How the engine worksRuns start on the events you describe, and nothing else.
Schedules and timers follow your calendar, time zones and holidays.
Follow an issue through its states, with timers like "2 business hours without a reply".
Behaviour tests and replays of past events run on every change.
Any event from the outside world can start a run: a message, a ticket, a webhook or a schedule. Agents act back through the same checked gateway.
Events in
Checked actions out
Real workspaces, shared outputs, lasting memory.
Each run gets its own sandbox, branch and files.
Project memory records where each fact came from.
Agents never hold your keys, code runs in a sandbox, and your words stay out of the logs.
The gateway adds credentials to each call. Agents never see them.
Secrets use envelope encryption with keys held in KMS.
Sandboxes run under gVisor, apart from the host.
Logs keep what happened, not what people wrote.
Hinzel is in early access. Tell us about your team and we will set up your first project with you.