AI you can ship.
Trust you can prove.
Place the Kirtonic Engine in front of your AI providers. Messages are scored against your policy, risky outputs are queued for review, and each decision is recorded in an append-only audit log.
A unified control plane for every AI interaction.
Define controls in plain English. The Engine categorises and enforces them across requests routed through it.
High-risk outputs are routed to designated reviewers. Each decision strengthens the classifier.
An append-only record of signals, decisions, and reviewer actions.
Reviewer decisions become labelled training data. Fine-tune a model on your organisation's policy.
Three steps. Live in fifteen minutes.
Type what should and shouldn't happen. The engine categorises each rule.
One URL change. Every request flows through the engine.
Reviewer decisions become labelled training data for your own classifier.
Govern shadow AI in every browser.
A free browser extension for Chrome, Edge and Brave. Install in two clicks.
Every prompt your team types into ChatGPT, Claude, Gemini, or Copilot is classified against your policy before it leaves the browser. Same engine, every browser.
Pick your browser → sign in → the download starts straight away.
Need the AI architecture behind your product? Deploy it in a prompt.
RAG, agents, model serving, fine-tuning pipelines, on AWS, GCP, or Azure.
Start free. Pay when you scale.
A small group of teams are running on Kirtonic during the pilot at no cost. Requests are reviewed manually so we can guide you through setup and learn what you need from the platform.
- 500 classifications a month
- Natural-language policy engine with reviewer queue
- Append-only audit log, 30-day retention
- Starter policy packs for PII, regulated advice, and injection
- Browser extension for Chrome, Edge, and Brave
- Works with major LLM providers (OpenAI, Anthropic, Bedrock, Gemini) and any OpenAI-compatible endpoint
- Support from the Kirtonic team
For teams that need higher classification volumes, multiple seats, identity-provider integration, or a deployment that meets specific data-residency and procurement requirements.
- Volume pricing on classifications and seats
- SSO and SCIM user provisioning
- Data Processing Addendum and Master Services Agreement
- Self-hosted in your VPC, or managed single-tenant SaaS
- UK, EU, or US data residency
- Availability targets agreed in the order form
- Named customer-success contact during onboarding
Enterprise-grade AI firewall,
deployed on your terms.
Two deployment models, designed around the data-sovereignty and operational requirements of regulated organisations.
Deploy in your own cloud
Deployed inside your environment. Your network, your keys, your control.
- Delivered asA signed container bundle. Docker Compose or Helm. Runs on AWS, Azure, GCP, or on-premises.
- Your dataNever leaves your network perimeter. Zero outbound dependencies in air-gapped mode.
- OnboardingGuided install with the standard bundle. Dedicated support included for the duration of the engagement.
We deploy and run it for you
Managed SaaS. Same isolation, none of the operational overhead.
- Delivered asA managed SaaS service. Your dedicated tenant, available on demand.
- Your dataPrivate database, private compute, isolated from every other tenant.
- OnboardingTenant provisioned on request. SSO and SCIM configured at activation.
Ship AI features aligned with the regulatory frameworks you operate under.
Sandbox tier includes twenty classifications a month so you can wire the engine in and watch the audit log fill up.