Simple, honest pricing

Free forever on Open Source. Pay for usage on Cloud. A flat annual license for on-prem Enterprise — never per-seat.

Monthly Annual Save 20%
Open Source
Free forever
Self-hosted. Apache 2.0. You own the deployment and the data — no call-home, no usage tracking.
  • All 13 pipeline stages
  • 20+ LLM providers
  • PHI/PII redaction (15 entity types)
  • Guardrails: regex + pluggable ML
  • Semantic cache + embeddings
  • Live model & pricing registry
  • Hosted dashboard
  • Managed billing
  • Priority support
Get on GitHub →
Cloud — Free
$0/month
We host it. Sign up and get an API key in minutes — no credit card required.
  • 10,000 requests / month
  • 1 project, hosted dashboard
  • Full 13-stage pipeline
  • Community support
  • SSO
  • FinOps showback export
Start Free →
Cloud — Starter
$49/month
For a small team running real production traffic through the gateway.
  • 250,000 requests / month
  • Unlimited projects
  • FinOps cost showback (CSV export)
  • A/B model experiments
  • Email support
  • SSO
Start Free Trial
Enterprise — On-Prem / VPC
Custom
Flat annual license, not usage billing. Install inside your own VPC or data center.
  • Unlimited requests
  • Full admin portal, self-hosted
  • HIPAA BAA available
  • Entra ID SSO on day one
  • Data residency enforcement
  • Custom PHI/PII patterns via YAML
  • Dedicated deployment engineer
  • Security review support
Contact Sales

Full feature comparison

Feature Open Source Cloud Enterprise

Frequently asked questions

Can I self-host for free? +
Yes. The full source is Apache 2.0 on GitHub. Deploy it anywhere — your own servers, Docker, or Kubernetes — free, with no restrictions on commercial use.
Does Cloud Growth include a HIPAA BAA? +
Not by default. A signed BAA is available on Enterprise (on-prem/VPC). If you're a covered entity handling PHI on Cloud, contact us — we evaluate this case by case.
What counts as a "request"? +
One call to /v1/chat/completions or /v1/embeddings. Streaming counts as one request. Cache hits count as one request but save the LLM cost entirely. We never bill per-token — you pay your LLM provider directly for the underlying model call.
Can I move from self-hosted to Cloud without downtime? +
Yes. The gateway is a drop-in proxy — point your base_url at the Cloud endpoint. Config and provider integrations import directly; no code changes in your applications.
Do you store my LLM responses? +
On Cloud, audit logs are stored encrypted at rest for compliance. The semantic cache stores a hash of the request/response, never raw PHI, and caching can be disabled for PHI traffic in config. On self-hosted or Enterprise, you control all storage.
What LLM providers are supported? +
20+, including OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, Google Gemini, Groq, Mistral, DeepSeek, Cohere, Fireworks, Together AI, Ollama, Perplexity, xAI, AI21, and more. Open a GitHub issue to request one.
Does it support Microsoft Entra ID? +
Yes — native Entra ID OIDC with RS256/ES256 JWT validation against Microsoft's JWKS endpoint, App Role enforcement, and multi-tenant support. Three auth modes: api_key, entra_id, or any (accept both, useful for zero-downtime migration). Included in every edition.
Is there a self-hosted admin portal? +
Yes — the Enterprise edition bundles the full governance portal (cost dashboards, request logs, governance rules, A/B experiments, team management) with the gateway engine, in a separate repository built for regulated industries that need everything on their own infrastructure. Deploys with docker compose up.

Talk to us

Evaluating the gateway for your organization? We're happy to run a proof of concept, answer compliance questions, or demo PHI redaction on your actual use case.

Contact information

Response Time
Within 1 business day for all inquiries
Open Source Issues
GitHub Issues

Why teams choose Enterprise

HIPAA BAA available
Dedicated Slack channel with your team
On-site or video onboarding
Custom PHI patterns for your workflows