Free IDE extension for risk-free vibe coding — keep secrets out of AI.
Detect
POST /api/canary/checkDeploy unique canary tokens as digital tripwires across your AI system. When an attacker successfully injects a prompt that causes the AI to leak or repeat protected content, the canary token appears in the output — alerting you to the breach instantly.
Complete these steps in order. The dashboard link opens the exact product workspace for this service.
Sign in, select the project you want to protect, then open Canary Network.
Insert unique canary tokens into protected contexts — agent instructions, system prompts, RAG documents, or email signatures. Each token is a unique, high-entropy string that looks like real data.
Use POST /api/canary/check from server-side code. Keep API keys out of browsers and client bundles.
Run one safe case and one expected-risk case. Confirm the result in Canary Network, then review Guard Logs before enabling the control for production traffic.
Traditional content filters look for known attack patterns, but novel attacks bypass them every time. Canary tokens are unique, high-entropy strings placed in sensitive contexts — system prompts, RAG documents, email signatures. If any output contains a canary token, you know for certain that protected context was leaked, even from an attack never seen before.
A step-by-step breakdown of how Canary Network protects your AI.
Insert unique canary tokens into protected contexts — agent instructions, system prompts, RAG documents, or email signatures. Each token is a unique, high-entropy string that looks like real data.
The system checks every AI output, tool call, API request, browser form submission, email, and memory write for canary token presence.
When a canary token is detected outside its intended context, the system blocks the leak, records the event with the exact location and content, and alerts your team.
Each leak event includes the session ID, location where the token was found, risk level, and redacted content — enabling rapid investigation of the breach source.
Cryptographically random, high-entropy canary tokens that are unique per deployment
Scans outputs, tool calls, emails, API requests, browser forms, and memory writes
Instant notification when a canary token is detected outside its intended context
Full context — session ID, location, risk level, and redacted content for each leak event
This service is implemented at the application route below. Use the tested SDK and framework guides for copy-paste integration examples.
POST /api/canary/check. API keys must stay server-side. See the REST API guide and related documentation before integrating.