Enterprise AI Control Plane
The control plane for enterprise AI, on your own infrastructure
MERIDVAR is software you install and operate. It discovers the AI in use across your organisation, records how each system and vendor was assessed and approved, refuses the requests your policy rejects before a model is called, and keeps the evidence. Nothing is sent to us.
- Self-hosted appliance
- Single tenant
- No telemetry
- Runs without internet access
Gateway admissions
Requests routed through this appliance.
- Provider calls
- 0
- Stored sample
- Masked
Why it matters
AI use has outrun AI governance
Employees use consumer AI. Developers wire in model APIs. Business units buy AI inside SaaS products. Agents act through MCP servers. Each decision was local, but the exposure is not.
- AI in use
- There is no single inventory of the systems, agents and vendors in use.
- Risk
- There is no shared view of the data involved, the autonomy granted or the decisions made about people.
- Data flows
- A prompt is an authorised request to an authorised service, with the data in the body.
- Spend
- Consumption is spread across teams, keys and providers.
- Accountability
- Nobody can show who approved what, on what basis and until when.
Source: IBM, Cost of a Data Breach Report 2025.
AI adoption is decentralised
AI arrives through employees, SaaS features, developer keys and vendor contracts, all at the same time. No single team sees all of it.
AI systems are becoming autonomous
Agents call tools, and MCP servers expose enterprise systems to models. A tool that can delete, pay or send data outside the organisation is a new kind of privileged access.
Accountability now has a schedule
EU AI Act transparency obligations have applied since 2 August 2026. Boards, auditors and regulators are asking the same question: how is AI governed here?
What it does
Five questions, answered in one record
Governance programmes start from declared inventories that are disconnected from the traffic. Security products start from the traffic and are disconnected from the decision. MERIDVAR keeps both in one place, on your network.
Stage 01
What AI is in use?
An AI inventory of systems observed in traffic or declared by people. Shadow AI discovery against your list of sanctioned providers. An agent registry, an MCP server inventory and endpoint agents.
Used by: Security and IT
AI inventory
Systems that this appliance observed in traffic or that a person declared.
An observation never becomes a declaration on its own. A person confirms it.
Stage 02
How risky is it?
Versioned questionnaires for AI systems, vendors, agents and MCP servers. Rules-based classification from Low to Critical, with a reason for every factor. Unknown is never read as safe.
Used by: Risk and privacy
Risk classification High
Computed by published rules from the completed baseline assessment.
This level requires a privacy review, a security review and an approved vendor before approval can be requested.
Stage 03
Who approved it, and until when?
Privacy and security reviews. Approval is decided by an administrator who did not request it, on a pinned basis, with conditions and an end date.
Used by: Privacy, legal and security
Approval Approved
Decided by an administrator who did not make the request.
- Requested by
- Operator, Customer Operations
- Decided by
- Administrator, Security
- Basis
- Classification, both reviews, vendor approval and inventory snapshot, pinned at the time of the request
- Conditions
- Block the delete tool on the linked MCP server
- Ends
- Twelve months from the decision
If anything in the basis has changed since the request, the approval is refused.
Stage 04
What is allowed through?
An admission gateway with a seventeen-stage pipeline: prompt DLP, threat detection, data residency, vendor requirements, budgets and approved models. A refused request never reaches the model provider.
Used by: Security, IT and finance
Gateway decision Refused
A request carrying a national identification number, sent from the console playground.
- Provider calls
- 0
- Stored sample
- Masked
The refusal names the rule that caused it. Nothing was sent, and nothing is billed.
Stage 05
Can we prove it?
An append-only, hash-chained evidence log. A status for each AI system and vendor that shows what has changed since each decision. An audit package for each subject.
Used by: Audit and compliance
Governance status Review required
The vendor’s classification changed after approval. Nothing is revoked automatically.
The audit package can be downloaded: every record about this system, as stored.
The product views on this site are illustrative, with invented example data. They show behaviour documented for the current release and are not screenshots of a customer environment.
How it works
The control plane runs where you run it
One appliance sits between your callers and your model providers. It runs five connected stages and keeps one record for every team.
What makes it different
The decision, the gate and the evidence, in one place that you operate
MERIDVAR is focused where a GRC suite or a security platform is broad. What sets it apart is its architecture.
No vendor cloud
There is no MERIDVAR control plane in our cloud and no telemetry. Prompt content is inspected, and evidence is kept, on your own infrastructure. No new data processor enters the path.
Governance tied to a gate that can refuse
The record of what was approved and the gateway that admits or refuses requests live in the same appliance and write to the same evidence log.
Decisions that visibly go stale
An approval keeps the basis on which it was granted. When an answer, a vendor decision or a classification changes, the approval reads “review required”, with the reasons.
Customer-controlled by design
Self-hosted, single-tenant and able to run offline, with signed licences and signed updates verified on the appliance.
Deployment modelHow MERIDVAR relates to the categories you already know
| Category | What it does well | Where MERIDVAR sits |
|---|---|---|
| AI governance platforms | Programme workflow, policy content and model risk | Governance tied to observed traffic and to a gate that can refuse, operated by the customer |
| AI security platforms | Broad AI threat coverage, from models to agents | Focused on admission control, with no vendor in the data path and the governance record included |
| SSE, CASB and SASE | Inline access control at network scale | Complementary: AI-specific governance, evaluated on your network |
| DSPM and data security | Finding and protecting data at rest | Complementary: governs decisions about AI systems and the prompt itself, before the call |
| LLM gateways and AI FinOps tools | Routing, caching and developer adoption | Spend by team and by model, recorded beside the security and governance decision |
| GRC platforms | Enterprise-wide risk and control libraries | The AI-specific record that a GRC programme can draw on |
Where a buyer is comfortable with cloud processing, an incumbent platform is often the right choice. MERIDVAR is built for the buyer who will not add a vendor’s cloud to the AI data path.
Who it is for
One control plane for every team that answers for AI
CISO
Security
Know what AI exists and what it is doing, and control the risk.
- Shadow AI discovered against your list of sanctioned providers
- Prompt DLP, threat detection and data residency enforced before the call
- Agents and MCP servers registered, reviewed and risk-rated
CIO and CTO
IT
Enable enterprise AI without losing control.
- An OpenAI-compatible endpoint that your teams and SDKs already use
- Approved model lists, with four providers supported for live dispatch
- A defined route to approval: assess, review and approve, with conditions
CFO and FinOps
Finance
Make AI financially accountable.
- A ledger of governed requests, with tokens, model and actual cost
- Spend by team and by model over time
- Team budgets with hard or soft caps, enforced at admission
DPO, compliance and legal
Privacy and compliance
Turn AI governance into auditable evidence.
- Every decision recorded with who made it, when and on what basis
- Approvals that read “review required” when the facts change
- An audit package for each AI system and each vendor
What the product looks like
Evidence and spend, read from the same appliance
MERIDVAR helps organisations operationalise governance controls and generate evidence relevant to the frameworks they answer to. Whether an organisation is compliant remains its own determination.
ISO/IEC 42001, Clause 8: operational controls Partial
Three of four observations are satisfied.
A control that the product cannot observe is reported as uncomputed, with the reason.
Control status computed from the running system
Control mappings ship for the EU AI Act, NIST AI RMF and ISO/IEC 42001. Records relevant to the GDPR, DORA, NIS2 and LGPD are kept as evidence, with no framework mapping.
Governed requests
Illustrative figures. Model names are placeholders.
A refused request makes no provider call, so it incurs no provider cost.
Cost recorded beside the decision
A ledger of governed requests, with tokens, model and actual cost, by team and by model. Team budgets are enforced at admission.
- Deployment
- A single-tenant appliance on your infrastructure. One appliance serves one organisation.
- Data handling
- No raw identifier is stored. Findings keep masked samples. Provider keys are never returned to the browser.
- Access control
- Administrator, operator and viewer roles, checked on the server. The approver is never the requester.
- Auditability
- An append-only, hash-chained log. A record edited on disk reads as an evidence mismatch.
Evidence
Measured, not promised
The product is held to the standard of evidence it applies to AI decisions. These results come from MERIDVAR’s own tests, not from an independent assessment.
Zero calls on refusal
23 gateway tests were run against a provider that logs every request it receives. Every refusal made zero calls, and every admission made exactly one.
No identifier in the clear
Prompt DLP findings keep masked samples. The raw identifier is not stored.
Offline updates
A signed incremental update was applied to a real appliance with no internet access. Licences were verified on the appliance.
Governance at scale
The status of 5,000 AI systems was computed in about 70 milliseconds.
Use cases
Where teams start
Most organisations begin with one urgent question. Each use case below has a vendor-neutral guide and a path into the product.
For investors
Building the control layer for the AI economy
Every enterprise that adopts AI needs to know what it runs, decide what is allowed, enforce that decision and prove it. That is a new control layer, with several budget owners and a record that grows more valuable over time.
See MERIDVAR refuse a request on your own network
A working session with the founder, on a host you control.
What happens next
- You send the formYour request goes straight to the founder.
- We reply by emailWe agree a time and what you want to see.
- We run a working sessionWe install the appliance with you, route a test request through it and read the evidence it leaves.