Secure your AI agents. Prove what they did.

estoppl intercepts every AI agent tool call, enforces guardrails, and produces a tamper-evident audit trail. Monitor and control every agent in your organization from one platform.

Request a DemoOpen Source Proxy

One platform to secure, monitor, and verify AI agent actions

AI agents call APIs, execute code, and move money. estoppl gives you control over what they can do — and proof of what they did.

Org-Wide Dashboard

Monitor every agent across your organization. Real-time event feeds, per-agent analytics, and alerting.

Guardrails & Policy

Allow lists, block lists, amount thresholds, rate limiting. Blocked calls never reach the upstream server.

Remote Kill Switch

Block a tool or shut down an agent from the dashboard. Every proxy picks up the policy change within seconds.

Cryptographic Audit Trail

Every tool call signed with Ed25519 and hash-chained. Tampering breaks the chain. Zero raw data stored.

Verification API

Upstream servers can verify that incoming agent calls are authorized and policy-compliant before processing.

Compliance & WORM Storage

Immutable audit records for regulated industries. SEC 17a-4, FINRA, EU AI Act. Export evidence packs on demand.

estoppl dashboard showing tool call events with ALLOW, BLOCK, and HUMAN_REQUIRED decisions

How it works

estoppl sits between your AI agents and the tools they call. Every action is intercepted, evaluated against policy, signed, and synced to the cloud for org-wide visibility and verification.

Your Agents
Agent 1
Agent 2
Agent 3
estoppl proxy
Policy evaluation·Ed25519 signing·Hash chaining
Allowed calls forwarded
MCP Server A
MCP Server B
MCP Server C
Blocked calls never reach upstream
Every event synced
estoppl Cloud
Dashboard
Alerting
Kill switch
WORM
Verification API

Read the full architecture →

Built for the teams deploying AI agents

Security Teams

“How do I prevent AI agents from exfiltrating data or calling unauthorized APIs?”

Allow lists restrict agents to approved tools only. Every blocked call is logged with the policy rule that triggered it. Tool poisoning and prompt injection attacks are stopped before they reach the server.

Compliance Teams

“How do I prove to auditors what our AI agents did — and didn't do?”

Ed25519 signed, hash-chained audit trail with zero data retention. Every event is tamper-evident and verifiable. Export reports or use WORM storage for SEC 17a-4.

Engineering Teams

“How do I add guardrails without changing code or slowing down development?”

One config change or run estoppl wrap to auto-detect and wrap all your MCP servers. Zero code modifications. Sub-millisecond overhead.

Platform Teams

“How do I verify that agents calling our APIs are authorized?”

The estoppl verification API lets upstream servers confirm an agent is known, the user is authorized, and the call passed policy — before processing the request.

Get started

estoppl Cloud

For organizations

Org-wide dashboard, verification API, remote kill switch, WORM storage, team access control, and compliance exports.

  • ✓Centralized monitoring across all agents
  • ✓Real-time alerting and kill switch
  • ✓Compliance-ready audit exports
  • ✓Verification API for upstream servers
Request a Demo
Open Source

For developers

Full proxy with guardrails, signed audit trail, and local dashboard. Apache 2.0. No account required.

npmnpm install -g estoppl
Homebrewbrew tap estoppl/tap && brew install estoppl
View on GitHub

Ready to secure your AI agents?

See how estoppl gives you visibility, guardrails, and provable audit trails across every agent in your organization.

Request a DemoOpen Source Proxy