Native records and receipts
Intents, logs, decisions, evidence, and effect receipts remain first-class records that can be inspected and replayed.
Prelaunch
AI-assisted work, governed from intent to effect.
Sidecat is a control plane for governed AI-assisted work. It keeps people, agents, authority, and evidence on the same accountable thread.
What exists today
Governance that stays attached to the work
Sidecat makes the path from a declared outcome to an external change inspectable without pretending the commercial product loop is open.
Governed work paths
Assignments, checkpoints, and handoffs stay bound to explicit work threads, so people and agents can resume the same accountable work.
Governed effect paths
Commits, pushes, service calls, and other external changes move through admitted operations with an actor, reason, and receipt.
Launch target · not live today
One clear path from purchase to proof
This is the launch target, not a statement of current availability. The paid stranger loop is not live.
-
Understand one published offer
See exactly what a nonzero payment buys before choosing it.
-
Establish bounded work
Create an account, bind the first work context, and approve the authority it may use.
-
Receive useful work and proof
Run one concrete outcome and inspect the records and receipts that show what happened.
Useful before launch
No signup form. Real work to inspect.
Pricing and the final launch call to action are not published yet. Until they are pinned down, follow the product’s development or read the source directly.