If it happened, it's on the record
Metorial follows every interaction from the first message to the final result. Your team always knows what happened and who was behind it.

- Every session, message, and tool call is recorded the moment it happens, with nothing to set up. The full sequence stays in session logs.
Operational visibility
Understand which tools were used, which policies checked them, and where issues appeared.
- Tool calls
- Errors
Activity Volume
Built for visibility
Catch issues and answer questions fast
Logs, monitoring, and search in one place. Go from a single trace to the full picture.
Session and message logs
A complete record of every session, message, and tool call your agents make, kept in session logs.
- Automatic session logs
- Full message and tool call history
- Detailed logging for errors and unexpected behavior
Monitoring and alerts
Watch activity over time and get notified the moment something looks off with monitors and alerts.
- Tool, error, and usage trends
- Alerts on unusual activity
- One operational view across agents
Search and review
Find past interactions with enterprise search, replay full sessions, and answer questions from security and leadership.
- Search across all agents
- Replay sessions end to end
- Connect findings to Protoguard
Every actor
Agents, people, and machines, all in one record
It does not matter who took an action. Everything lands in one place security, IT, and leadership can trust.
AI agents
Every action ties back to the specific agent that triggered it and the user behind it with agent identification.
- One identity per agent
- Tied to the person running it
People
What your team does through MCP and Portals is recorded, on their real identity via Magic MCP.
- Actions taken in the API
- Agents connected through Portals
- Tied to a real person, not a shared login
Machines
Service accounts and API calls are captured the same way as everything else.
- Service account activity
- API and SDK calls
- Nothing happens off the record
Frequently asked questions
Common questions about Tracing.
What is the difference between tracing, observability, audit logs, and session logs?
Tracing is the visibility layer that ties them together. Observability is real-time operational health, audit logs are the who-did-what accountability record, and session logs are the detailed end-to-end record of a single connection.Can I see what people and machine users do, not just AI agents?
Yes. Actions from people in the dashboard, agents calling tools, and service accounts on the API are all recorded in one history through audit logs.How does Tracing tie an action back to a specific agent and person?
Each agent has its own identity. Every call ties back to the agent that made it and the person behind it. See agent identification.How long is trace data kept, and can I control it?
Retention is configurable per data type to match your compliance requirements. See data retention.Can I export traces for auditors and compliance?
Yes. Records can be filtered, exported, and retained on your terms. You can answer auditors without holding more than you need. See audit logs.Does Tracing capture security threats like prompt injection?
Yes. Threats detected by Protoguard appear in the record and can trigger alerts. See prompt injection monitoring.


