What it does
Runs produce portable receipts that can be replayed without the model, credentials or original store; approval gates remain reserved for people or policy code.
Agent tooling
A language and runtime for budgeted agent programs whose declared decisions can be answered by Jev and replayed later.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
Runs produce portable receipts that can be replayed without the model, credentials or original store; approval gates remain reserved for people or policy code.
Try ALGAL when an AI task in your project needs to wait for someone's approval. A developer can write the steps and limits in its language. Pause a task, review it, and then continue. ALGAL keeps a record of completed work so you can replay it without asking the AI again. Approval still belongs to people or written rules.