Quickstart
Scaffold a Tarski app, run verification to check that its contract holds, and deploy the same app to the cloud in about five minutes.
Every scaffolded Tarski app passes verification from its first minute. The scaffold is a working curriculum app that ships with its own fixtures. This page walks through the local loop first, then the cloud path.
The local loop
Scaffold an app
tarski scaffold --pattern minimal my-app cd my-appThe
minimalpattern is domain-neutral. It contains a commented ontology, a stub mapper, a non-vacuous happy-path fixture, and a manifest. Use--pattern defaultfor a fuller curriculum app with a declared query, decision gating, and richer fixtures. Other patterns:sensor(fallible extraction),decision(LLM proposal gating),chat(agent interface),end-user-auth(verified identity and policy-gated access).Prove the contract
tarski verifyverifyreplays every fixture on a clean database against the deterministic evaluator and runs these checks over the replay: golden expected outputs, invariants, provenance, replay determinism, and redaction. It always ends with a verdict that is both human- and machine-readable. A green verdict means the declared contract holds.Replay a timeline
tarski replay fixtures/happy-path.jsonl --isolatedA fixture is an observation timeline (
.jsonl) plus an expected world (.expected.json). Replay shows you the derived model those observations produce, and--isolatedleaves your workspace’s default lineage untouched.Start the dev loop
tarski devThe dev loop watches your app’s text assets (
.taontology, Rhai mappers, and fixtures) and hot-reloads when they change. There is no compile step. Apps are directories of text files that thetarskibinary interprets.Inspect in Studio
tarski studioStudio shows observations, derived facts, intents, effects, and provenance. Provenance explains why each derived fact exists by tracing it back to the observations that support it. See Studio.
What verify checked
The scaffold’s fixture states a timeline of observations and the exact derived world that
must result from it. tarski verify ran that fixture against the same evaluator that runs
in production, so there is no separate test-harness translation layer that can drift. When
you or your coding agent change a rule, the fixture either still holds or names what
changed.
The cloud path
When you are ready to run hosted, deploy the same app to Tarski Cloud:
tarski cloud login --email you@company.com
tarski cloud whoami
tarski cloud select --project my-project
tarski cloud deploy
tarski cloud observe --kind example.created --payload '{"id":"example-1"}'
Login uses a code-only email flow. Tarski emails you a sign-in code, which you can enter
interactively, pass with --code, or supply through --code-command for agent-driven
workflows. A denied cloud select exits nonzero and leaves your previous scope untouched,
so a chained select && deploy cannot deploy to a stale target.
Where next
- Project structure: what each scaffolded file does.
- The Tarski language: relations, rules, invariants, namespaces.
- Scenarios & fixtures: the proof model in depth.
- Authoring with agents: how to point Claude Code, Codex, or Devin at the contract.