The Tiresias Platform
One platform, three pillars
of AI agent security
Identity. Detection. Enforcement. Tiresias-ZT governs your agents' access to their memory and MCP backends from inside your own infrastructure - seeing threats, never taking your data.
How the platform works together
Agent requests flow through identity verification, real-time monitoring, and gateway enforcement - end to end.
Better Together
Each product is powerful on its own, but the real magic happens when they work as an integrated platform. Identity feeds monitoring. Monitoring informs enforcement. Enforcement validates identity.
SoulAuth establishes who an agent is and what it can do. Every action starts with verified identity.
SoulWatch monitors what agents actually do against what they should do. Behavioral baselines, anomaly signals, quarantine.
SoulGate enforces deny-by-default policy at the request layer. Bad requests never reach your backend.
Three products. One mission.
Secure every agent, every action, every API call.
SoulAuth
Private BetaAgent Identity & Zero-Trust Access
Every AI agent gets a durable cryptographic identity. Every action is evaluated against deny-by-default policy. No standing permissions, no over-provisioned tokens, no blind trust.
Learn moreKey capabilities
- Cryptographic SoulKey identities
- Zero-trust policy decision point
- Capability-token introspection
- Policy-as-code
- RBAC: Platform → MSSP → Tenant → Agent
- OIDC single sign-on
SoulWatch
Private BetaBehavioral Anomaly Detection
Behavioral detection for your agent fleet. Baselines, anomaly signals, and automated quarantine - watching how agents behave at the request layer, not reading your data at rest.
Learn moreKey capabilities
- Per-agent behavioral baselines
- Anomaly detection across fleets
- Automated quarantine
- Usage metering + budget governor
- Monitor-first rollout mode
- SoulKey-signed audit of every action
SoulGate
Private BetaDeny-by-Default Policy Enforcement
The enforcement point between your agents and their memory / MCP backends. Deny-by-default decisions and injection screening at the request layer - with no LLM in the decision path.
Learn moreKey capabilities
- Reverse proxy for memory / MCP backends
- Deny-by-default request evaluation
- Prompt-injection screening
- Sovereignty-gated reads
- Rate limiting & cost governor
- Full request audit logging
Ready to secure your agents?
Tiresias is in private beta. Request a design-partner slot for early access to the full platform.