The single thermostat over your observability stack.
Sentry tells you errors. Datadog tells you traces. Better Stack tells you uptime. xcp-platform tells you whether your AI-governed system is shipping responsibly — by reading from the tools you already pay for, and adding the AI-governance signals nobody else collects.
The same AI-governed infrastructure layer that Xlooop itself runs on — now available to engineering teams shipping AI-governed systems.
Methodology brief available under NDA on request.
Your observability is fragmented. Your governance is not observable at all.
Modern infrastructure teams run three to five observability tools at once. Sentry for errors, Datadog or Honeycomb for traces, Better Stack or Pingdom for uptime, Cloudflare or CloudWatch for logs, plus whatever the database vendor ships. Each tool has its own dashboard, its own alerting rules, its own pricing tier, and its own answer to a different question.
None of those tools answer the question that actually matters when you ship AI-governed systems: are we governed well enough to deploy this?
That question lives in places your APM does not look. Did teams finish their work properly. Did closing ceremonies happen. Are operator decisions resolved on time. Are verifiers honest. These are first-class operational signals for AI-governed software — and there is no off-the-shelf product that collects them.
xcp-platform was built because we needed those signals to ship Xlooop responsibly. We are making it available to other teams that ship AI-governed systems and want one operator-grade surface instead of five engineer-grade ones.
What makes xcp-platform different from the tools in your stack.
Built for operators, not engineers.
Sentry shows you 453 errors this week. xcp-platform shows you "your AI-governance readiness is at 78% — below the 85% target. Three teams skipped their wave-close review last sprint. Here is the list." Different audience. Different actionability. The operator number, not the engineer number.
AI-governance signals nobody else collects.
Discipline metrics, decision-state SLA, deferred-items pile depth, methodology changelog drift, ceremony depth across waves, skill-invocation diversity, verifier honesty. These are the AI-governance signals that catch regressions months before a Sentry alert would fire — because the regressions show up in process before they show up in errors.
We aggregate. We do not replace.
xcp-platform sits ABOVE your existing observability stack through a small adapter pattern. Sentry stays Sentry. Better Stack stays Better Stack. We read their APIs and add what they cannot see. Swap any vendor for its open-source equivalent (GlitchTip for Sentry, Uptime Kuma for Better Stack) by changing one environment variable. No vendor lock-in.
Infrastructure-tier — not a second SaaS subscription.
Because we aggregate instead of replace, xcp-platform's marginal cost is the cost of running our adapter and dashboard — not the cost of a second SaaS contract. Private beta runs on a single VPS for most teams. The architecture is honest, so the savings story is honest.
Patented multi-agent collaboration model.
The role-based collaboration model is covered by US Patent 11,604,641 (granted March 14, 2023 · 24 claims · 2 independent). Claim 1 covers multi-user collaborative roles plus chat-bot integration — i.e. how multiple AI agents work alongside humans with role-scoped authority. The governance engines (risk-based gates, evidence ledger, delegation) are built on top of the patent — not themselves patented. European decision due October 2026. Adevi IP Holding Company Pty Ltd is the assignee; ADEVI Pty Ltd (trading as Xlooop) is the operating company. Read the granted patent on Google Patents.
Where xcp-platform sits in your observability chain.
Think of it like a thermostat above your thermometers. The thermometers measure. The thermostat decides.
┌──────────────────────────────────────────────────────────────────┐ │ YOUR EXISTING TOOLS (we read; we do not replace) │ │ ┌──────────┐ ┌────────────┐ ┌────────────┐ ┌────────────────┐ │ │ │ Sentry │ │Better Stack│ │ Datadog / │ │ Workers Logs │ │ │ │/GlitchTip│ │/Uptime Kuma│ │ Honeycomb │ │ / CloudWatch │ │ │ └─────┬────┘ └─────┬──────┘ └─────┬──────┘ └────────┬───────┘ │ │ │ │ │ │ │ │ └────────────┴──────────────┴─────────────────┘ │ │ │ │ │ (vendor APIs) │ │ ▼ │ │ ┌────────────────────────────────────────────────────────────┐ │ │ │ xcp-platform — adapter layer + aggregation + decisioning │ │ │ │ ──────────────────────────────────────────────────────── │ │ │ │ Adapters: Sentry · Uptime · Neon · Workers Logs · custom │ │ │ │ AI-governance signals: M1-M7 discipline metrics + │ │ │ │ composite readiness signal │ │ │ │ Output: ONE number + 4-tile operator surface │ │ │ └────────────────────────────────────────────────────────────┘ │ │ │ │ │ ▼ │ │ ┌────────────────────────────────────────────────────────────┐ │ │ │ Operator dashboard — xcp-control-plane │ │ │ │ "Are we ready to ship?" — one number, drill-in on click │ │ │ └────────────────────────────────────────────────────────────┘ │ └──────────────────────────────────────────────────────────────────┘
On mobile: think of three layers stacked. Bottom layer = the observability tools you already pay for. Middle layer = xcp-platform adapters reading their APIs and adding AI-governance signals. Top layer = your operator dashboard surfacing one composite readiness number.
Trend over time, not snapshots.
Composite readiness charted across 30 days, with the 85% target line and coloured bands where the line crossed below target (red) or above (green). Proves your team measures discipline over time, not just on the day of a regulator visit.
Built for teams that ship AI-governed systems.
- YOU, IFyour team has 2+ AI agents (Claude · Codex · Cursor · etc.) writing code into shared repos, and you need to see when they drift, regress, or skip closing rituals.
- YOU, IFyou maintain compliance-relevant infrastructure (financial services, healthcare, AU privacy regimes, GDPR-bound EU SaaS) and need governance evidence on demand for regulators or insurers.
- YOU, IFyou have already paid for Sentry / Datadog / Honeycomb but still find yourself opening five tabs to answer "how are we doing this week?"
- YOU, IFyour team is small (single operator, 1-10 engineers) and cannot justify a Datadog Enterprise contract — but the cost of a governance miss is real.
- YOU, IFa Sentry alert at 2am yesterday should have fired at 2am the day before — but did not, because closing-rituals slipped through and nobody noticed for 24 hours.
Seven AI-governance signals nobody else collects — plus one composite.
The questions on the left are what you ask. The codes in brackets are how engineers refer to them internally. Both reach the same number on your operator dashboard.
Did teams finish their work properly?
Percentage of multi-file commits that cite the closing ceremony (retrospective + lessons + audit). When this drops, technical debt is piling up silently. Target ≥ 80%.
Are AI actions measurable?
Percentage of AI-skill invocations that emit verifier-readable telemetry. When this drops, you lose audit-trail provenance. Target ≥ 80%.
Did teams run the closing ceremonies?
Percentage of wave-close commits with full ceremony (retrospective + lessons-learned + audit pack). When this drops, ship velocity looks fine while quality degrades.
Are you using your full capability catalogue?
Unique skills invoked / canonical catalogue size. When this drops, teams are over-relying on a small subset of tools — the rest atrophies.
Is your telemetry complete?
Percentage of skill invocations with complete event chains (start + checkpoints + close). When chains break, you cannot prove what AI did for whom.
Are operator decisions resolved on time?
Percentage of operator-decision states resolved within 72 hours. When this drops, the team is shipping past unmade decisions — accumulating governance debt.
Are your verifiers honest?
Percentage of verifiers passing the no-inflation check (do they report failures truthfully?). When verifiers start cheating, every other metric becomes meaningless.
Are we ready to ship?
Weighted composite of M1-M7 with operator-tunable thresholds. The single number that decides whether the next release goes out, gets held, or triggers a wave-close ceremony first.
The infrastructure layer underneath Xlooop.
xcp-platform is the AI-governed infrastructure layer that Xlooop itself runs on. Same governance discipline. Same audit-grade evidence chain. Same operator-tested foundation. Two products, two buyers, one shared infrastructure.
Xlooop · AI cockpit
Sold to regulated SMBs — accounting, building inspection, owner-operator firms 5-50 staff. A team of specialist AI agents led by a dedicated Chief of Staff. Buyer is the managing partner or principal.
See Xlooop →xcp-platform · AI-governed infrastructure
Sold to engineering teams shipping AI-governed systems — SREs, heads of infra, technical co-founders. The thermostat over your existing observability stack. Buyer is the operator owning ship-readiness.
This page.Every vendor in our adapter list has an open-source equivalent.
The adapter pattern means every external tool in your stack can be swapped for its open-source equivalent without changing the xcp-platform dashboard or the integration code. Document the swap path in your operator runbook; we publish ours.
Run a pilot on a Mac launchd cron at no marginal cost. Scale to a $5/mo VPS for redundancy. Multi-region at $10-15/mo. The architecture is honest, so the price is honest.
Infrastructure-tier — not a second SaaS subscription.
xcp-platform is in controlled private beta as of 2026-Q2. Pilot tier is the active offer. Standard and Enterprise tiers follow once pilot signal lands.
Pilot
Single team · controlled environment · signed pilot agreement
First 3 months free during private beta
Pilot tier is the active offer through 2026-Q2.Standard
Production deployment · multi-region observability adapter · regulator-ready exports
Infrastructure-tier · talk to us
Pricing reflects operating cost, not SaaS rent extraction.Enterprise
Multi-tenant · SSO / SCIM · custom adapters · AU/EU/UK/US residency
Per-organisation · talk to us
Includes dedicated CSM + named SLA + region-pinning.We would rather lose your evaluation now than lose your trust later.
- —
Not an alternative to your APM
We read from Sentry / Datadog / Honeycomb. We do not replace them. If you do not already have observability, install Sentry first.
- —
Not for teams without AI agents in the codebase
If your shipping process is humans-only and AI is just for writing emails, xcp-platform is over-engineered for you. Pick this up when AI starts producing code that ships.
- —
Not a compliance certification
We surface governance signals. We do not issue SOC 2 attestations. Pair with Drata / Vanta / Secureframe for certification; use us for the operating layer underneath.
Want to see your stack’s governance signal?
If you ship AI-governed software and want to see what M1-M7 looks like for your team, get in touch. Pilot tier is the active offer through 2026-Q2.
Methodology brief + architecture-decision records available under NDA on request.
xcp-platform is part of the Xlooop family of products. Entity: ADEVI PTY LTD (trading as Xlooop) · ABN 80 627 916 698 · ACN 627 916 698 · Australia (NSW jurisdiction). Multi-jurisdictional terms apply per Terms of Service; data-residency posture per Data Sovereignty; AI disclosure per Privacy Policy.
Three doors. Same destination: a firm whose AI work is provable.
Pick the one that fits where you are today — none commit you to anything.
Or simply email hello@xlooop.com — one human reads every message.