Five moves. Everything we make serves one of them.

Find the AI tools and agents you build, buy, or didn't know about.

• EndpointsFinds the AI already running on employee laptops.
• Browser extensionWeb AI apps, personal and corporate accounts, seen.
• Discovery connectorsCopilot, Foundry, Bedrock, Vertex, Agentforce, MSP platforms.
• Agent registryThe system of record for agents and MCP tools.
• Signals and domainsSignals start runs; domains let agents call your API without a connection string.

See every call, token and dollar your instrumented agents spend, run by run.

• ObserveInstruments the agents you build outside Waxell.
• TelemetrySpans, tokens, latency, cost, without altering behavior.
• ExecutionsEvery run as a record you can open and export.
• BudgetsSpend per agent, per customer, per feature.
• CLIYour whole agent stack from the terminal.

Decide allow, hold, redact, or block while the agent acts.

• Policies50+ categories, from PII to spend caps to egress.
• MCP GatewaySecures every tool call in flight.
• ConnectGoverns the third-party agents already on your team.
• Governance testingProve enforcement works before it has to.
• Skills and pluginsScan every Claude skill and plugin before it is installed.

Run the agents you build inside the plane, with a kill switch that works.

• RuntimeGates each step of the high-risk workflows you build on Waxell.
• CLIPush, tail and schedule from the terminal.

Evidence per decision, mapped to the frameworks.

• ComplianceEU AI Act, NIST AI RMF, ISO 42001, OWASP LLM Top 10, from one record.
• Audit trailThe policy, the decision, the person, the reason.