# Fairvisor

> Open-source edge enforcement engine for API rate limiting, LLM cost control, and agentic loop protection. Deployed at the edge as a reverse proxy or decision service. Sub-millisecond enforcement, declarative policy, self-hosted or SaaS.

## Key facts

- Decision latency: p50 < 100μs, p99 < 1ms, p99.9 < 5ms
- Bot patterns: 1,335 across 7 categories (44 AI crawlers)
- Kill-switch propagation: < 10 seconds to all edges
- Token counting: prompt tokens before request, completion tokens during SSE streaming
- Integrations: Nginx (auth_request), Envoy (ext_authz), Kong, Traefik, AWS/GCP/Azure API Gateway
- Compatible LLM APIs: OpenAI, Anthropic, Azure OpenAI, Google Gemini, vLLM, Ollama, LiteLLM
- Pricing: Open Source (free, self-hosted), Pro ($299/mo), Scale ($499/mo), Enterprise (custom)
- License: Open-source core, SaaS control plane

## Product

- [How It Works](https://fairvisor.com/llm/how-it-works/index.md)
- [Quickstart](https://fairvisor.com/llm/quickstart/index.md)
- [Pricing](https://fairvisor.com/llm/pricing/index.md)
- [Security](https://fairvisor.com/llm/security/index.md)
- [Open Source](http://github.com/fairvisor/edge)
- [Enterprise](https://fairvisor.com/llm/enterprise/index.md)

## Solutions by Role

- [For AI Teams](https://fairvisor.com/llm/for/ai-teams/index.md): Token budgets, loop detection, cost controls for LLM agents in production
- [For Platform Engineering](https://fairvisor.com/llm/for/platform-engineering/index.md): Policy-as-config, GitOps-native, Kubernetes-ready rate limiting infrastructure
- [For FinOps](https://fairvisor.com/llm/for/finops/index.md): LLM cost attribution by tenant, team, and endpoint; real-time budget enforcement
- [For SRE](https://fairvisor.com/llm/for/sre/index.md): Sub-millisecond enforcement, graceful degradation, SLO alerting, incident runbooks
- [For Compliance](https://fairvisor.com/llm/for/compliance/index.md): Immutable audit logs, RBAC with MFA, SOC 2 control mapping

## Solutions by Industry

- [For LLM Providers](https://fairvisor.com/llm/for/llm-hosters/index.md): Anti-extraction controls, identity-aware enforcement, forensics at the inference layer
- [For API-First SaaS](https://fairvisor.com/llm/for/api-platforms/index.md): Per-tenant JWT-based limits, noisy neighbor protection, tiered plan enforcement
- [For FinTech](https://fairvisor.com/llm/for/fintech/index.md): Per-partner quotas, ASN-type policies, deterministic failure, audit trail
- [For AdTech & Media](https://fairvisor.com/llm/for/adtech/index.md): ASN-aware rate policies, Tor/hosting tagging, burst shaping
- [For Crypto & Web3](https://fairvisor.com/llm/for/crypto/index.md): IP-tiered limits, paid vs free enforcement, abuse shaping for public APIs
- [For Content Sites](https://fairvisor.com/llm/for/content-sites/index.md): 44 AI crawler patterns, staged rate limiting, shadow mode analytics

## Comparisons

- [Fairvisor vs Kong](https://fairvisor.com/llm/compare/kong/index.md)
- [Fairvisor vs Cloudflare](https://fairvisor.com/llm/compare/cloudflare/index.md)
- [Fairvisor vs LiteLLM](https://fairvisor.com/llm/compare/litellm/index.md)
- [Fairvisor vs robots.txt](https://fairvisor.com/llm/compare/robots-txt/index.md)
- [Fairvisor vs AWS API Gateway](https://fairvisor.com/llm/compare/aws-api-gateway/index.md)
- [Fairvisor vs Azure API Management](https://fairvisor.com/llm/compare/azure-api-management/index.md)
- [Fairvisor vs GCP API Gateway](https://fairvisor.com/llm/compare/gcp-api-gateway/index.md)
- [Fairvisor vs Nginx](https://fairvisor.com/llm/compare/nginx-rate-limiting/index.md)
- [Fairvisor vs Envoy](https://fairvisor.com/llm/compare/envoy-rate-limiting/index.md)

## All LLM-friendly pages

https://fairvisor.com/llm/