Get in Touch
AI content toolsfrom Opace

Developer tools · Version 0.1.0

Opace AI Content Integrity CLI and Local Service

Inspect files from a terminal or connect an approved local client to the authenticated loopback service. Results follow the same frozen methods and receipt contract.

Evidence sheetBrowser-only
SourceExact text and protected factsHeld in memory for this run
SignalsNamed checks with limitsNo universal authorship score
Watermark scanKnown keysReal SynthID mathematics; Anthropic production keys stay private
Evidence summary

01 / Current release state

Publication candidate 0.1.0

Local package checks passed

The dependency-closed TypeScript packages and deterministic Python distributions passed local independent submission review. Nothing has been published to npm, PyPI or a container registry.

02 / Product scope

Terminal inspection, typed clients and an authenticated loopback service

The command-line route returns bounded text, JSON or receipt output for scripts and continuous integration. Approved local applications can use the typed client and service contract for managed jobs. The service stays on loopback, requires bearer authentication and keeps raw tokens out of the process environment.

Terminal workflow

Inspect supported text from scripts and continuous integration with bounded JSON, text or receipt output.

Typed client

Use the frozen local-service operations through a fail-closed TypeScript client with bounded SSE handling.

Loopback service

Run authenticated local jobs without exposing a public listener or retaining raw bearer tokens in the process environment.

03 / Release evidence

Installation, protocol lifecycle and no-egress checks

Clean installation

The TypeScript package set and Python wheel and source archive install in fresh, locked environments.

Protocol and lifecycle

Error envelopes, job transitions, cancellation, shutdown and token handling have adversarial regression coverage.

No-egress baseline

The model-free route runs without content-bearing network calls, provider credentials or telemetry.

04 / Requirements and limits

Before running the packages or local service

Match the declared Node and Python versions, keep the listener on loopback and issue separate run and administration credentials. The model-free baseline needs no detector provider. Semantic models, paid-provider calls and container distribution remain separate, unavailable routes until their own evidence and approvals exist.

  1. Node and Python versions must match the package metadata for the selected client or service.
  2. Keep the service on loopback and use separate run and administration tokens.
  3. Semantic model, paid-provider and Docker distribution routes remain unavailable until their separate approvals are complete.

05 / Frequently asked questions

CLI and local service questions.

Does the CLI require an internet connection?

The model-free inspection route works locally. It does not contact a detector provider or send the inspected text to Opace.

What does the local service expose?

The Python service binds to loopback, requires bearer authentication and provides the frozen health, capability, inspection, job and receipt contracts.

Does this include semantic AI detection?

No. Semantic model and corpus work remains unconfigured pending separate data, model, legal and benchmark approval. The capability state reports that limit rather than inventing a result.

Where should I report a problem?

Use the Opace contact page and include the product version, platform version, expected behaviour and the smallest safe reproduction. Do not include draft text, authentication tokens, personal data or customer information. Contact Opace.

Submission candidate

Use the checker now. Return for the release.

The software and submission materials are being checked as one release set. Until approval, this page records the product scope without presenting an unapproved download or store listing.