Designed for regulated industries

Run Corelayer in your environment, with controls your security, risk, AI governance, and audit teams can verify.

Deployment · BYOC, on-prem, and SaaS options availablePII masking onRetention: none
Your environment
Your systems
payments-db
read-only role
Splunk
logs and alerts
GitLab
payments-api, settlement-service
Corelayer
Agents
investigation and monitoring
Context engine
stored in your account
On-prem gateway
connectivity to on-prem systems
Inference
Your LLM gateway
your keys, your provider
PII masking
Payment declined for Name, card Card, receipt to Email

Sensitive data doesn't have to leave your environment

Start in our cloud or deploy into your own account or on-prem infrastructure.

Settings/Deployment model
SaaS
Corelayer-managed cloud (dedicated option available)
Bring your own cloud
Your AWS, Azure, or Google Cloud account
On-premises
Your infrastructure, including Red Hat OpenShift
Bring your own cloudHealthy
CloudAWSAzureGoogle Cloud
Account482193746021payments-prod
Regioneu-west-1
InferenceCorelayer defaultsYour LLM gateway

PII redaction layer

Configurable PII masking for your most sensitive data.

FieldRaw valueWhat the agent sees
customer_nameMaria LindqvistName
emailm.lindqvist@example.seEmail
card_number4111 2203 9981 4417Card
auth_tokeneyJhbGciOiJIUzI1NiJ9.e30Secret
client_ip10.0.8.1210.0.8.12
podsettlement-worker-7d4b8csettlement-worker-7d4b8c
Settings/PII masking
SecretsAlways
Personal information
Financial data
Network addresses
System identifiers

Run inference through your LLM gateway

You have the option to route all LLM calls through your enterprise gateway, so all traffic uses your centralized guardrails and controls.

Corelayer agents
Investigations and monitoring
Your LLM gateway
llm-gw.internal.bank
PII and secrets filteringApproved models onlyRate and spend limitsRequest audit log
Anthropic
OpenAI
AWS Bedrock
Google Vertex AI

Access that matches your org chart

SSO, role-based access control, and SCIM provisioning, with customer-managed credentials for every connected system.

RolesSSO · OktaSCIM synced
RoleView issuesRun investigationsApprove actionsManage integrations
Admin✓✓✓✓
On-call engineer✓✓✓–
Engineer✓✓––
Auditor✓–––

Every agent action is controlled by deterministic guardrails

You configure access policies and human approval gates, and everything has an audit trail.

What Corelayer can accessSet by the access you grant
EKS prod-eu-1Read only
payments-db replicaRead only
SplunkRead only
GitLab merge requestsRead and write
Autonomy with constraintsYou choose the policy per action
Restart settlement-workerSuggestRestart with approvalAutomatically restart
Roll back a bad deploySuggestOpen MR with approvalAutomatically open MR
Audit trailEvery action is recorded with its evidence
09:44:52Agent proposed rolling back payments-api config3 citations
09:46:10Approved by Shipra J., payments-oncallCHG-20931
09:46:14Agent opened merge request !482GitLab
09:58:03Agent verified error rate back to baseline2 citations

Trusted by the most regulated industries

Corelayer works with top 10 US banks, Fortune 500 fintech companies, and healthcare companies that handle PHI.

Banking
Including top 10 US banks
Fintech
Including Fortune 500 fintech companies
Healthcare
Including companies protecting patient data
AICPA SOC 2
SOC 2 Type II compliant
Independently audited controls, with reports and documentation in our trust center.
Visit the trust center →

Frequently asked questions

Can Corelayer run fully on-premises?

Yes. Corelayer runs on customer-managed infrastructure, including on Red Hat OpenShift.

Which clouds does bring your own cloud support?

AWS, Azure, and Google Cloud.

Does Corelayer store our production data?

Corelayer has zero retention of production data by default.

Is our data used to train models?

No customer data is used for model training.

Can we use our own LLM provider or gateway?

Yes. You can route inference through your own model provider or LLM gateway with your own credentials.

How does PII masking work?

You configure which sensitive fields are masked. Masking can be run before data is sent to the LLM.

Is Corelayer SOC 2 Type II compliant?

Yes. Reports are available through the trust center.

Experience Corelayer in your environment.