InspectAgents - AI Agent Testing & Safety Platform

InspectAgents

API & Integration Documentation

Everything you need to integrate InspectAgents into your AI testing workflow. All endpoints are free with no authentication required.

Quick Start

Get started in 30 seconds. No signup, no API keys needed.

# Search for prompt injection failures
curl -X POST https://inspectagents.com/api/mcp/ \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"my-agent","version":"1.0.0"}}}'

REST API

POST /api/agent-feedback

Submit feedback, corrections, or new AI incident reports.

Rate limit: 10 requests/minute per IP

curl -X POST https://inspectagents.com/api/agent-feedback \
  -H "Content-Type: application/json" \
  -d '{
    "type": "incident",
    "message": "AI chatbot leaked internal data",
    "source": "Claude",
    "url": "https://example.com/article"
  }'

Parameters

Field	Type	Required	Description
`type`	string	Yes	feedback, correction, incident, or suggestion
`message`	string	Yes	Content (max 5000 chars)
`source`	string	No	Agent identifier (e.g. "Claude", "ChatGPT")
`url`	string	No	Reference URL

GET /api/agent-feedback

Returns self-documenting API docs as JSON. Useful for agent discovery.

MCP Server (Model Context Protocol)

InspectAgents exposes a Streamable HTTP MCP server for direct AI agent integration.

Endpoint

POST https://inspectagents.com/api/mcp/

⚠️ Trailing slash required. Without it, you get a 308 redirect.

⚠️ Accept header required: application/json, text/event-stream

Rate limit: 30 requests/minute per IP

Available Tools

Tool	Description
`search_failures`	Search the AI failures database by query, category, severity, year
`get_failure`	Get full details of a single failure by ID
`assess_risk`	Run the AI risk assessment quiz
`get_checklist`	Retrieve the 67-point testing checklist
`submit_feedback`	Submit feedback or new incidents

Connection Flow

Step 1: Initialize

{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"my-agent","version":"1.0.0"}}}

Step 2: List Tools

{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}

Step 3: Call a Tool

{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"search_failures","arguments":{"query":"prompt injection","limit":5}}}

Resources (Read-Only)

URI	Description
`inspectagents://failures/all`	Complete failures database as JSON
`inspectagents://checklist/full`	Full 67-point checklist

Error Handling

All API endpoints return structured JSON errors.

Status	Cause	Response
400	Invalid request body or missing fields	JSON with error details and required fields
404	Unknown API endpoint	JSON with available endpoints
405	Wrong HTTP method	JSON with allowed methods
406	Missing Accept header (MCP only)	Include `Accept: application/json, text/event-stream`
429	Rate limit exceeded	JSON with rate limit info; `X-RateLimit-*` headers
500	Internal server error	JSON-RPC error with message

Webhooks

Webhook support for real-time notifications is planned. Currently, you can poll the MCP server for updates using search_failures with date filters.

To be notified when webhooks launch, email hello@inspectagents.com.

Discovery & Integration Files

File	Purpose
/llms.txt	LLM-optimized site overview
/llms-full.txt	Full context with statistics and examples
/api/openapi.json	OpenAPI 3.1 specification
/.well-known/mcp.json	MCP server discovery
/.well-known/mcp/server-card.json	MCP server card
/.well-known/agent-card.json	A2A Agent Card
/.well-known/ai-plugin.json	ChatGPT Plugin manifest
/.well-known/agent-skills/index.json	Agent Skills index (v0.2.0)
/.well-known/api-catalog	RFC 9727 API catalog
/pricing.md	Machine-readable pricing
/index.md	Markdown homepage fallback

Rate Limits

Endpoint	Limit	Window
POST /api/mcp/	30 requests	per minute per IP
POST /api/agent-feedback	10 requests	per minute per IP
POST /api/subscribe	5 requests	per minute per IP

Rate limit info is returned in X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers.