Overview

Start building with the Fireraven developer platform by choosing the product path that matches your job:

FireGuard protects production agents with input and output guardrails.
Redraven evaluates AI systems with automated red-team tests and compliance reports.

Protect an agent with FireGuard

Create a conversation, check user input, call your model, then check the output before returning it.

Protect AI coding agents

Add FireGuard checks to Windsurf, Devin, Cursor, Claude Code, and Copilot Studio.

Evaluate a model with Redraven

Run a red-team test from your code while keeping model keys in your environment.

Explore the API reference

Pick FireGuard or Redraven endpoint pages with schemas, examples, and request builders.

First API Calls

Use FireGuard when you need a production request path that blocks unsafe user input before it reaches your model:

curl -X POST \
  "https://api.fireraven.ai/public/fireguard/v1/conversation?project_id=$FIREGUARD_PROJECT_ID" \
  -H "X-Api-Key: $FIREGUARD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"Production chat"}'

Use Redraven when you need to evaluate an existing AI system and export the results:

curl -sS \
  -H "X-API-Key: $REDRAVEN_API_KEY" \
  -H "X-Organization-Id: $REDRAVEN_ORGANIZATION_ID" \
  "$REDRAVEN_BASE_URL/api/v1/tests/$TEST_ID/results"

Choose Your Path

Goal	Start here
Add guardrails to a live chat or agent	FireGuard quickstart
Protect AI coding agents	FireGuard Agent Hooks
Export monitoring data from FireGuard	Monitoring API reference
Run Redraven evaluations from Python	Redraven Python SDK
Pull Redraven metrics and reports over HTTP	Redraven API authentication