Tailor this page to you
You’re mandated to adopt AI — but your data can’t leave the perimeter, only certain models are approved, and most AI testing SaaS is disqualified on day one. Karate runs entirely inside your network, on your own model, with a token bill you can cap.
One Docker container on your infrastructure. Bring any model — cloud, approved, or fully local via Ollama. No outbound calls, no telemetry, no hosted control plane — and the deterministic core runs with zero AI at all.
What stays inside your walls
Your data
No egress, no telemetry — ever
Your models
BYO-LLM, including fully local
Your network
One Docker container, air-gap ready
Your budget
Token cost capped · ~$0 replay
Local-first by design — not a deployment option
No outbound calls.·No telemetry.·No hosted control plane.
Every component runs inside your perimeter. We don’t know what you’re testing — unless you tell us.
The business case
The board wants AI everywhere. You own the blast radius if it leaks data or runs up an uncapped bill. This is how you say yes, safely.
The whole stack runs inside your perimeter — no egress, no telemetry, air-gap ready. The deterministic core runs with no AI at all, so security can buy it before a single model is turned on.
BYO-LLM: your approved Azure OpenAI / Copilot models, or open-weight models you host yourself (Llama, Qwen, Gemma) via Ollama. Your keys, never captured by a vendor.
Token-frugal by design, with ~$0 deterministic replay and usage on every report. Scale by adding a container — not by buying another row of per-seat licenses.
Our data can’t leave the country, only certain models are approved, and most AI testing SaaS is disqualified by security on day one.
— What regulated-enterprise security teams tell us
You feel this when…
Under the hood
What leaves your network: nothing. The entire pipeline — browser, agent, model — runs where you put the container.
Cloud AI testing SaaS
Karate — self-hosted
# everything runs where you put the container
docker run -p 8080:8080 \
-e LLM_ENDPOINT=http://ollama.internal:11434 \ # your network
-e LLM_MODEL=gemma3 \ # open-weight, local
karatelabs/karate-agent
# no outbound calls · no telemetry · no keys leave your VPCHow we keep it inside your walls
The whole pipeline ships as a single Docker image you run on your infrastructure — CI, cloud, or fully air-gapped bare metal. The unit of scale is a box, not a seat.
Self-hosted AI testingPoint it at your approved cloud model or an open-weight model you host. DOM-first means small, local models work — you don’t need a frontier vision model, or a vendor’s token markup.
Bring your own LLMSAML/OIDC SSO, RBAC, audit logs, and a reproducible no-AI report — the controls procurement and security ask for. Local-first, with nothing phoning home.
Security & complianceUse cases
Data-residency and audit obligations met by design — the stack runs inside your network with reproducible evidence.
Approved modelsAlready standardized on an approved model? Point Karate at it — no new vendor in the data path, no second AI bill.
Air-gappedRun the whole pipeline on bare metal with a local model — zero external calls, even during heavy regression.
Deploy the container in your environment, point it at an approved or open-weight model, and watch the token meter — nothing leaves your network.