Security designed for AI that searches and acts
Layered safeguards, granular governance, and infrastructure that meets the bar of the most demanding security teams — without slowing down the people using Referent every day.
Enterprise security built for AI that takes action
Referent is engineered with layered safeguards so your data stays governed, your models stay accountable, and your team stays in control, no matter how much work the assistant ships for you.
Your data, your rules, end to end
Decide what Referent can see, retain, and act on. Granular controls keep sensitive information protected at every layer of the stack.
Live federated retrieval
Referent fetches information in real time from your sources. Sensitive content never has to be copied or stored on our infrastructure.
Indexing controls
Choose exactly which spaces, channels, folders, and document types Referent is allowed to read, share, and act on.
Zero data retention
Our LLM provider agreements enforce zero retention. Your prompts and content are never stored or used to train any model.
A platform that holds up under audit
Resilient infrastructure, encrypted transport, and full observability so security and IT teams have the evidence they need.
Fault-tolerant infrastructure
Built without single points of failure so search and assistant workloads stay available, even during partial outages.
TLS & domain security
All traffic is encrypted end to end with TLS 1.2+, 2048-bit RSA certificates, and SHA-256 signing.
Single sign-on (SSO)
SAML-based SSO and SCIM provisioning let admins integrate with the identity provider they already trust.
Audit logging
Every connection, query, and admin action is logged with device and IP metadata for review and export.
Single-tenant architecture
Each Referent workspace runs inside its own isolated environment, dedicated to one company.
Workspace activity log
Monitor usage, surface trends, and prove compliance with full visibility into how teams interact with the assistant.
Stay in control of the models behind every answer
Pick the model, route the traffic, and own the keys. Referent adapts to your governance, not the other way around.
Choose your LLM
Power Referent with leading frontier models from OpenAI, Anthropic, Google, or open-source providers, depending on your performance and compliance needs.
Bring your own API key
Use your own provider account to retain full visibility, rate limits, billing, and audit logs over every model interaction.
Bring your own cloud
Deploy Referent inside your AWS, Azure, or Google Cloud environment so data, networking, and compliance never leave your perimeter.
Need a deeper review? Our team can walk through architecture diagrams, data flows, and compliance documentation with your security stakeholders.