EnfinitOSEnfinitOS
Developers

Build governed execution into every screen you ship.

28 SDKs across every substrate EnfinitOS governs — DOOH, CTV, mobile, AR / glasses / HUD / volumetric / hologram, gaming, automotive, smart home, wearables, audio, messaging, robotics, drone, satellite, aviation, maritime. The auditor / verifier trio is open source under MIT — no operator has to trust EnfinitOS as a vendor to verify the proof we issue.

28
SDKs published
23
Substrates covered
15
Languages
MIT
Open-source verifier
How to read this page

Three honest signals on every SDK card

We label every SDK with its real shipping status so you do not find out the hard way. Substrate readiness is tracked openly in the public readiness matrix, separately from the per-SDK status shown below.

Open source · MIT

Published on GitHub. The auditor / verifier trio lives in public so any operator, regulator, court, or third-party compliance tool can fork it, fuzz it, and verify our proof packs without trusting us.

Production-ready scaffold

Typed, tested, documented, and wired to the EnfinitOS platform endpoints that exist today. Vendor-side SDK integrations (Broadsign / VIOOH / DJI / Tizen / Alexa / Twilio / Stripe etc.) land per-customer at pilot integration time — those bring the renderer / transport / exchange-specific code; the EnfinitOS half is ready.

Alpha

Design and package scaffolding complete; the substrate- specific runtime layer (Unity/Unreal/Godot engine plugins, watchOS/Wear OS Tiles) is not yet implemented. Use these as a contract preview, not a deployable build. Reach out before relying on them.

EnfinitOS Governance Copilot

Deterministic execution. Grounded intelligence. Human-approved governance.

The rights gate, policy engine, proof issuance, metering, and settlement remain fully deterministic — every decision is signed, reproducible, and reconcilable. The Copilot layer sits on top of that evidence and translates it into operator-readable explanations, anomaly summaries, and draft proposals. The deterministic engine still decides; the Copilot only assists.

Every Copilot call is grounded in the underlying audit-log entry, logged through the platform's Observability layer, and reconcilable byte-for-byte from the evidence hash. We follow the NIST AI Risk Management Framework and the EU AI Act human-in-the-loop posture by construction — not by retrofit.

Live · 15 features

  • Why was this blocked?. Plain-English explanation of why a delivery, offer, or rights operation was refused by the deterministic gate. Reads the audit-log entry, the basis, and the relevant policy, and translates them into operator-readable prose. Starter+ tier.
  • Anomaly explanation. Reads a slice of the event stream, runs classical-ML anomaly detection, and emits an operator-readable summary of what looks off and why. The decision about which signals are anomalous is deterministic; AI only writes the prose. Growth+ tier.
  • Proof pack summary. Customer-safe executive summary of a sealed proof pack. Reads the receipts, metering, settlement, and exceptions; emits a defensible explanation suitable for sharing with advertisers / auditors / regulators. Growth+ tier.
  • Policy drafting. Plain-English → structured policy JSON. Drafts only; deterministic engine still executes the final approved policy after human review. Scale+ tier.
  • Operator Copilot. Natural-language interface to your tenant's audit stream + dashboards. 'Show me every rights conflict for this estate.' Read-only; never mutates state. Scale+ tier.
  • Reconciliation explanation. Read-only narrative over metering + settlement gaps. Identifies likely root causes ('312 screens offline 14:02-16:19'); does not propose credit amounts. Scale+ tier.
  • SDK developer assistant. Helps developers integrate the EnfinitOS SDKs. Generates example code, explains API errors, suggests fixes, runs conformance checks against documented contracts. Growth+ tier.
  • Delivery risk prediction. Pre-launch risk assistant. Given a campaign brief + target estate, predicts which placements will be blocked or under-deliver and proposes mitigations. Scale+ tier.
  • Pilot brief → policy draft. Reads a pilot agreement or scope document and drafts a structured policy JSON. Subset of contract-ingestion, useful pre-launch. Drafts only — operator reviews and approves before the engine sees the policy. Growth+ tier.
  • Anomaly morning brief. Scheduled daily digest of anomalies across your tenant. Runs at a configurable time each morning; emails the operator a prose summary of what looks off and the suggested follow-ups. Growth+ tier.
  • Substrate-safety attestations. For high-stakes substrates (DRONE, ROBOTICS, AUTOMOTIVE), generates a plain-English safety narrative from the deterministic policy gate. Designed to be filed with regulators / safety boards. Scale+ tier.
  • Multi-language operator UI. Translation layer over explain-block + operator-copilot for non-English operators. Wraps existing AI outputs rather than running separate inference. Growth+ tier.
  • Onboarding copilot. First-week hand-holding for new operators. 'What should I do next?' guidance grounded in their current sandbox state. Conversion-focused, not deep-feature. Starter+ tier.
  • Consent decay assistant. Reads upcoming consent.expiresAt and drafts re-consent campaigns ahead of the deadline. GDPR-aligned; the deterministic side already tracks expiry, this writes the operator-readable summary + suggested action. Growth+ tier.
  • Cost / tier-fit advisor. Reads a tenant's usage and suggests whether their current tier is the right fit. Conversion-driving on the upgrade side; trust-building on the downgrade side. Growth+ tier.

Pre-launch · 2 scaffolded

Sitting on the hardening pipeline, ready to switch on as operator feedback shapes the prompts:

  • Auditor-facing proof-pack Q&A
  • Brand / Auditor Copilot

Post-launch · 2 scaffolded

Enterprise-only surfaces with higher-stakes review requirements. Held back until an enterprise customer names a specific need:

  • Contract ingestion
  • Reconciliation proposal
Try the Copilot · live demo

Pick a feature, pick a scenario. See the deterministic evidence + the AI translation side-by-side.

Plain-English explanation of why a deterministic mutation was refused. The model is grounded in the audit-log entry; the prose is its translation.

Rate-limited to 3 calls per hour per IP across both demos. Sign up to run unlimited calls against your own audit log.
SDK catalog

28 SDKs · 8 categories · 15 languages

Every render SDK composes the substrate-agnostic Renderer Core — composition, not inheritance. Drone composes Robotics, not the other way round. Adapter-mostly SDKs (Satellite, Aviation, Maritime) translate vendor vocabulary into the EnfinitOS plane rather than wrapping native UI.

Operator & brand

What operators and advertisers consume to integrate with EnfinitOS — the React component library, the operator CLI, and read-only REST clients for the brand side of every campaign.

Renderer core

The substrate-agnostic foundation every render SDK wraps. Owns the rights / policy / proof plane; substrate-specific SDKs compose it, they do not inherit from it.

Environment

Render SDKs for screens embedded in the consumer's environment — vehicle dashboards, smart-home displays, wearables, ambient audio, and outbound messaging.

API reference

The HTTP API behind every SDK

All 28 SDKs are thin clients over the same governed HTTP surface — rights, offers, challenges, proof, metering, settlement, consent, drills, audit export. The OpenAPI 3.1 spec is generated from the same definitions the platform validates against; live reference docs publish alongside the April 2027 platform launch.

  • OpenAPI 3.1 schema, single source of truth
  • Try-it-out in browser against a sandboxed tenant
  • Code samples in TypeScript, Python, Go, Java, Ruby, Rust
Sandbox · live

A real sandbox you can run a contract through

A scoped tenant, real rights handshake, real signed proof chain, and an in-browser auditor — running now in your browser tab. The sandbox issues packs in the same envelope.v1 format the production platform uses; every signature, every chain link, every metering projection is the real thing.

  • Provision a tenant in one click
  • Watch synthetic delivery events seal into the chain
  • Verify Ed25519 signatures locally — no server trust required
Trust model

Don't trust us. Verify.

The auditor / verifier trio is open source under MIT so any regulator, auditor, court, or compliance tool can verify EnfinitOS proof packs without taking the vendor's word for any of it. Ed25519 signatures over canonical JSON; SHA-256 hash chain; metering re-projected from the same source events the platform settles against.

Signed

Every proof pack is signed by EnfinitOS with a long-lived Ed25519 key whose public half is published. Tamper with any byte and the signature fails.

Chained

Each proof pack references the SHA-256 hash of the previous one for the same tenant. Drop one and the chain breaks visibly. The verifier walks the chain end to end.

Re-projected

The verifier re-projects metering from the same source events the platform settles against and re-runs settlement reconciliation. Output is bit-identical or the pack fails.

Get in touch

Founder-led, no sales filter

EnfinitOS is a founder-led pre-launch platform — built deliberately small, with enterprise-grade engineering and operational discipline. Every email lands with the founder until the team grows. No ticket queue, no sales filter, no SDR cadence.

Security disclosures

Coordinated disclosure to [email protected]. Scope + GPG key publish alongside the April 2027 platform launch.