Skip to content
CONTROLLED AI · PEOPLE + AGENTS

AI for your people.Control for your agents.

Sluis gives employees a secure AI workspace and sends AI-agent requests through one governed gateway. One policy, one audit trail, any device, with SSO.

See how Sluis works
01 / API

Three APIs, every model.

Drop-in for the OpenAI and Anthropic (Messages) APIs: swap your base_url and you are behind the lock.

02 / INSPECT

Personal data stays inside.

Detection and reversible pseudonymization before anything leaves; the key never exits your perimeter.

03 / SEAL

Proof per request.

A sealed, hash-chained ledger records every passage, ready for your auditor.

04 / RESOLVE

Best model, still inside policy.

Call sluis/auto or a use-case alias; Sluis refreshes managed routes daily from Artificial Analysis and resolves the target inside your residency policy.

One prompt, three views.

The same request: as your user types it, as the model receives it, and as your app gets it back.

What your user types

“Draft a payment plan for Anna de Vries, anna.devries@mail.nl, citizen ID 123456782.”

What the model sees

“Draft a payment plan for «PERSON_1», «EMAIL_1», «ID_1».”

What your app gets back

The full answer, with Anna back in place, restored mid-stream, no visible delay.

your apprequestSluis gatejohn@acme.com«EMAIL_1»«EMAIL_1»masks PIImodel«EMAIL_1»«EMAIL_1»token onlySluis gate«EMAIL_1»john@acme.comjohn@acme.comrestoresresponseyour appyour app · requestSluis masksjohn@acme.com«EMAIL_1»«EMAIL_1»PII stays insidemodel«EMAIL_1»«EMAIL_1»token onlySluis restores«EMAIL_1»john@acme.comjohn@acme.cominside the streamyour app · restored

Every request passes three gates.

No request skips a gate. None leaves without a record.

your app01inspect · dlp02route · residency03seal · ledgermistral · eu-parisscaleway · eu-amsopenai · us-eastallowed · policydeepseek · cn403 · not in policyyour app · request01inspect · dlp02route · residency03seal · ledgermistral · eu-parisscaleway · eu-amsopenai · us-eastallowed · policydeepseek · cn403 · not in policy

Inspect

The gate closes behind you.

60+ detectors scan every prompt before it leaves; what they find is reversibly pseudonymized.

Route

The water level equalizes.

You decide per key: which models, which region, whose ownership; a failed provider fails over automatically.

Seal

The passage goes in the ledger.

Every request gets a hash-chained entry: verifiable offline, exportable, erasable per person (GDPR Art. 17, or any right-to-delete duty).

A data center address is only half the answer.

An address says where a server stands, not who owns it. Legal reach follows ownership: a provider's parent company can be compelled under laws like the CLOUD Act.

  • Region and owner

    Where a model runs and who owns it are two different risks; Sluis labels both axes on every model in the catalog.

  • EU by default

    API residency defaults to EU-only. You can additionally require EU-owned providers, and explicitly allow other regions when your policy permits.

  • Per workload

    One key allows a frontier model for general drafting; another stays locked to the strictest route for anything personal.

residency.policy · class: PHIenforced
EU · France
mistral/mistral-large-latest
primary
EU · France
scaleway/llama-3.3-70b
fallback
EU · region
vertex/gemini-3.5-pro
allowed
United States
xai/grok-4.3
blocked · policy
MCP / TOOL GATEWAY

Govern the tools your AI agents can call.

Every MCP tool call follows the same policy as a model request, with deny-by-default access and a sealed audit record.

Deny-by-default grantsResidency-routed toolsSealed in the same ledger
See the MCP gateway →
tool.crm.readLIVE
01grant.customer.readALLOW
02anna@bedrijf.nl → «EMAIL_1»MASK
03route.eu-ownedEU
04sha256:8f21…b40aSEALED
Enterprise · optional

Need it on your own hardware? Run Sluis Edge.

The whole data plane as a single binary on your own iron: same gates, same ledger, your perimeter.

On your hardwareData never transits SluisOne console
See Sluis Edge →
AGENT HARNESS

Claude Code, Codex and Cursor, governed without leaving the laptop.

The agent, its files and its built-in tools stay on the developer's machine. Its model requests and MCP tools travel on dedicated agent routes with a purpose-isolated key, while your organisation's Claude, Codex and Cursor subscriptions stay encrypted in the gateway.

Agent stays localSubscriptions held gateway-sideSigned MCP plugins
See the Agent Harness →
SIMPLE TARIFF

One price: 10% on model usage.

Your provider's list price, plus a flat 10%. Pay per token from the first one, no subscription.

See pricing

Keep the models your teams already use. Sluis applies policy, data protection and audit evidence before every request leaves.

Chat · Skills

From a question to a governed answer.

Use Chat with any approved model. Personal data is pseudonymized before it leaves, reusable Skills standardise the workflow, and every request is sealed in the audit ledger.

Open Sluis Chat →
Sluis ChatEU
YOU

Summarise the renewal contract for the board. Keep names inside our perimeter.

PII SEALEDEU ROUTELEDGER VERIFIED
SLUIS

Summary ready. Names stayed inside your perimeter and the answer used an approved model.

seal · 8f21…b40a

Put the lock in between.

Get a key