What works today

OpenAdapt compiles demonstrated GUI workflows into deterministic, locally executable programs. Healthy runs make no model calls. When an interface drifts, the runtime can re-resolve a target deterministically or use an explicitly configured model tier; it records any repair and halts when the configured identity, postcondition, effect, or policy checks cannot verify the run.

That is the canonical product description. A bounded hosted recorder has passed a live-provider record-to-compile qualification for public, non-regulated browser targets. The complete paid account-to-run lifecycle remains a Beta launch candidate pending production acceptance. That does not mean every recorded workflow is safe or every backend is production-ready.

Maturity labels

Label	Meaning
Beta	Implemented and exercised end to end in the stated environment, but still requires workload-specific validation. Public delivery may have a separate production-qualification gate.
Experimental	Implemented and tested in a bounded environment; interfaces or operating assumptions may change.
Research spike	Demonstrates an adapter or mechanism, usually with mocks or an analog environment; not a validated integration.
Prototype	Code or a deployment scaffold exists, but required production controls are incomplete.
Target-state	Designed or planned; do not treat it as available.
Deprecated	Kept for compatibility or history; do not build new integrations on it.

No surface is labelled Stable yet.

Integrated product matrix

Surface	Status	What is actually demonstrated	Important boundary
openadapt installer and openadapt flow dispatcher	Beta	Installs the compiler and exposes the unified command surface.	The standalone openadapt-flow package remains the canonical engine and may be installed directly.
Record -> compile -> lint -> certify -> replay -> report on a browser	Beta / reference path	Runs end to end in CI against the bundled app; the browser path is also the only backend exercised end to end against a real third-party app.	A clean run is not automatically safe. Identity coverage, risk classification, postconditions, and effect contracts must be audited per bundle.
Deterministic target re-resolution and saved heal diffs	Beta on the reference path	Theme, movement, and rename drift are covered by the bundled drift matrix.	Scale/reflow and tenant-specific state can still halt. The base bundle is not silently promoted; save and review a healed bundle explicitly.
lint and certify	Beta	Report coverage gaps and refuse bundles that violate the selected policy.	Certification is opt-in and only enforces what the policy names. An uncertified bundle remains runnable with replay.
Fail-closed run admission gate	Experimental	The shipped gate checks certification, identity/effect coverage, approval fallback, encryption, and manifest integrity before executing.	Development escape hatches exist, and passing the gate does not validate the backend or prove a workflow safe.
Identity verification	Experimental, opt-in by step coverage	Structured identity works on armed browser steps; UIA and pixel/OCR tiers are implemented and adversarially tested.	Unarmed clicks have no identity check. Pure-pixel ambiguity intentionally over-halts, and coverage varies by workflow.
System-of-record effect verification	Experimental	REST, FHIR, and document-hash verifiers run in the live replay path and halt on refuted or indeterminate declared effects.	The compiler does not infer effects. A deployment must author effects and configure the matching verifier; otherwise screen checks remain the oracle.
teach halt-to-correction loop	Experimental	The deterministic reference inducer covers the optional-dialog correction class behind regression and canary gates.	Arbitrary UI corrections are not generally learned. Unsafe or underdetermined revisions are refused.
Workflow-program IR, branches, loops, and multi-trace induction	Experimental	Implemented against synthetic fixtures, including refusal for underdetermined programs.	No real recording exercises the full Phase-2 program path yet.
Windows UIA backend	Experimental	A local Windows-on-ARM VM demonstrated record -> compile -> replay and database-ground-truth checking; adapter tests run against mocks.	Evidence is small-N and not a native-x86 or broad application study.
Native macOS backend	Target-state	macOS capture/input primitives exist for the remote-display analog harness.	There is no validated native macOS AX backend integration today.
RDP backend	Research spike	The adapter conforms to the backend protocol and is mock/offline tested.	No published live RDP validation establishes latency, coordinate, input, identity, or effect behavior.
Citrix backend	Research spike	A pixel-only remote-display analog was exercised against a VM window.	It is not a validated Citrix/ICA/HDX integration. Real latency, compression, DPI, lock screens, input acceptance, and clinical identity behavior remain unmeasured.
Desktop authoring GUI and tray	Experimental	Native artifacts for the targeted platforms and their tag-gated release pipeline have been built and qualified; the surfaces follow the record -> compile -> replay -> teach loop.	No complete native prerelease is published yet. Signing, permissions, updater behavior, and the full operator lifecycle still require platform validation.
Hosted CLI connectivity	Beta / launch component	login, exact-hash artifact preparation/upload, one-time runtime validation, bound replacement activation, and report-break connect the local engine to control-plane endpoints.	Upload requires destination policy and an approved sanitized derivative; checkout never bypasses an egress refusal. Public paid acceptance remains pending.
Artifact sanitation and local review	Beta / launch gate	The sanitized-derivative pipeline inventories, transforms, rescans, manifests, hashes, and supports local review/approval.	The raw original remains sensitive; unknown or unresolved content is refused; runtime observations can reintroduce PHI.
Cross-engine hosted validation	Beta / launch gate	validate-hosted binds an approved recording and bundle, compiler provenance, strict lint, policy certification, derived risk class, and successful replay report to a one-time Cloud challenge.	It is operator self-attestation signed with the ingest token, not an independently observed certification. Exact deployment policy, risk-class, and deployed compiler-version allowlists still apply.
Hosted browser recorder	Beta / bounded launch component	A real Modal session produced PNG frames and retained input evidence, assembled a compileable recording, finalized one workflow idempotently, enforced resource limits, and cleaned up ephemeral qualification data. The qualified worker used openadapt-flow 1.7.3.	Explicitly initiated, public-HTTPS, non-regulated authoring only. Raw observations remain private inside the declared hosted boundary; this does not qualify paid checkout or managed replay.
Hosted dashboard/control plane	Beta launch candidate	Authentication, organizations, exact-hash bundle ingest, immutable run admission, browser runner orchestration, structural reports, replacement activation, billing, and metering form the candidate launch lifecycle.	Local governed authoring and repair remain available. Production must explicitly use live dependencies; mock mode remains visibly synthetic development behavior. The current paid lifecycle has not passed clean-account acceptance.
Hosted execution	Beta launch candidate	The implementation routes Stripe Checkout through onboarding to managed browser execution; the runner verifies exact admitted bundle bytes and callbacks.	This is not yet a public availability claim. The configured offer covers browser workflows, not Windows, RDP, or Citrix by implication. Checkout does not create an SLA or certification. The full paid production lifecycle remains pending.
Air-gapped on-prem package	Experimental	A local queue, systemd unit, minimized hash-chained audit log, and air-gap checks are provided.	Full-disk encryption and operational hardening remain operator/deployment responsibilities.
BYOC / customer-controlled execution	Experimental / scoped deployment	A connector and customer-storage contract support keeping the data plane inside an approved customer boundary.	Qualify the actual substrate and destination; PHI-bearing runtime data must not cross into a shared boundary.
Hosted desktop runner	Experimental	A Windows runner contract and implementation path exist.	Not included in the browser launch candidate and not a validated broad desktop/Citrix service.

Before using consequential data

Treat the matrix as a starting point, not an authorization. For each workflow:

1. Inspect lint findings and identity-armed coverage.
2. Review every risk classification and write-shaped action.
3. Declare effects and configure an independent verifier where a real system of record exists.
4. Certify against a workload-specific policy.
5. Exercise expected drift and failure cases with repeated trials.
6. Review the PHI boundary, storage encryption, logs, model egress, updates, rollback, and operator approval design.
7. If an artifact crosses a boundary, inspect its sanitation manifest and approve the exact derivative hash required by policy.

The complete, evidence-linked failure inventory is maintained in openadapt-flow/docs/LIMITS.md.