One API.
Every AI Model.
Access 400+ models from OpenAI, Anthropic, Google, DeepSeek through a single endpoint. Transparent routing, semantic caching, zero BYOK fees.
Join 0+ developers on the waitlist. $50 free credits at launch.
Everything OpenRouter has.
Plus what it doesn't.
Transparent routing, semantic caching, compliance, self-hosted models, plugins, observability, and a full admin portal.
Transparent routing with explainable decisions
Every routing decision returned in response headers. See exactly why each request went to a specific provider. Train routing with feedback. Visual flow in dashboard.
ExclusiveSemantic caching
Three-tier: exact match, semantic similarity, provider prompt cache passthrough. Save 40-80% on repeated queries.
ExclusiveZero-fee BYOK
Bring your own keys with 0% commission. Always. OpenRouter charges 5% after 1M requests.
One-click compliance
HIPAA, SOC 2, GDPR, EU AI Act. PII redaction, prompt injection detection, toxicity filtering.
ExclusiveAgent-native MCP gateway
Per-workflow budgets, automatic model degradation, tool call filtering, session management, and multi-step workflow tracing with cost per node.
ExclusiveSelf-hosted models
Ollama, vLLM, TGI, or any OpenAI-compatible endpoint. Auto-discovery, health checks, cloud fallback.
ExclusivePlugins & presets
Web search, PDF parsing, response healing, code interpreter. Config-as-code presets with versioning.
Broadcast observability
Datadog, Langfuse, LangSmith, Braintrust, OTEL, and 8 more. Per-key filtering, sampling, privacy mode.
Every team. Every workflow.
Whether you're a solo hacker or a 500-person engineering org, AllRoutes scales with you.
AI-Powered Products
Ship faster with auto-fallbacks, cost optimization, and observability built in.
- Automatic failover across providers
- Per-model cost tracking & budgets
- A/B test models without code changes
AI Agents & Workflows
MCP-native gateway with session management and multi-step tracing.
- Per-workflow budgets & rate limits
- Tool call filtering & safety tiers
- Cost-per-node tracing
Enterprise & Compliance
HIPAA, SOC 2, GDPR ready. PII redaction. EU data residency. Full audit trail.
- One-click compliance profiles
- SSO/SCIM & role-based access
- On-premise deployment option
Cost Optimization
Semantic caching saves 40-80%. Smart routing picks the cheapest model that meets your quality bar.
- 3-tier semantic cache
- Intelligent model selection
- Credits that never expire
Developer Tools
Build internal tools with the best model for each task, without managing multiple API keys.
- Single endpoint for all models
- Interactive playground
- SDKs for 7+ languages
Self-Hosted & Hybrid
Run Ollama or vLLM locally. AllRoutes auto-discovers your models and falls back to cloud.
- Auto-discovery of local models
- Health monitoring & cloud fallback
- Keep sensitive data on-prem
AllRoutes vs OpenRouter
Same where it matters. Better where it counts.
AllRoutes Better
OpenRouter
Three steps. Sixty seconds.
No complex setup. No vendor lock-in. Start for free.
Create account
Sign up with Google, GitHub, or email. Get $50 free credits instantly. No credit card required.
Get API key
Scope to specific models, set budgets and rate limits, assign guardrails, or just use smart defaults.
Ship it
Point your OpenAI SDK to api.allroutes.ai/v1. Every model works instantly with auto fallbacks.
Pay as you go. Upgrade for more.
No markup. No credit purchase fees. No expiring credits. Subscriptions reduce platform fee and unlock features.
- 10K requests/mo
- 4% platform fee
- 0% BYOK fee
- 1 custom endpoint
- Basic guardrails
- 7-day logs
- Unlimited requests
- 2.5% platform fee
- Semantic caching
- 15 routing rules
- All plugins & presets
- Broadcast (2 dest.)
- 30-day logs
- Unlimited requests
- 2.5% platform fee
- PII redaction
- Compliance profiles
- EU data residency
- 50 team members
- Audit logs
- Unlimited everything
- Negotiable fee
- SAML SSO / SCIM
- Dedicated SLA
- VPC / on-premise
- Custom compliance
- 24/7 support
Two lines to switch.
Zero lock-in.
Your OpenAI SDK code works instantly. SDKs for every language. Interactive docs. CLI for power users.
Drop-in replacement
Change base URL + API key. Every model from every provider works.
Side-by-side playground
Test models head-to-head. Tune params, see costs live, share via link.
Interactive docs
Every endpoint has Try It. Request builder generates code in any language.
SDKs & CLI
Python, Node, Go, Ruby, PHP, Java, Rust. CLI for key management & testing.
# pip install openai from openai import OpenAI client = OpenAI( base_url="https://api.allroutes.ai/v1", api_key="allroutes_sk_..." ) r = client.chat.completions.create( model="anthropic/claude-sonnet-4.6", messages=[{"role": "user", "content": "Hello!"}], extra_body={"models": [ "openai/gpt-5.4", "google/gemini-3.1-pro" ]} )
What builders are saying
From developers who've seen the early preview.
"Finally a gateway that tells me WHY it chose a specific model. The transparent routing headers are a game-changer for debugging."
"Switched from OpenRouter. Zero BYOK fees and the semantic caching already saved us 60% on repeated embedding queries."
"The compliance profiles are exactly what we needed for HIPAA. PII redaction + audit logging without building it ourselves."
Common questions
Yes. AllRoutes is fully OpenAI SDK compatible. Change your base URL to api.allroutes.ai/v1 and your API key. Your existing code works instantly with every model from every provider.
Pay-as-you-go with credits. No markup on model pricing. Free tier: 4% platform fee. Pro ($29/mo) and above: 2.5%. BYOK is always 0%. Credits never expire.
Bring Your Own Key means you use your own provider API keys through AllRoutes. You pay providers directly, we charge 0% commission. You still get routing, caching, analytics, and fallbacks.
Same model catalog and compatibility. But AllRoutes adds transparent routing, semantic caching (40-80% savings), compliance profiles, PII redaction, self-hosted model support, an admin portal, and charges less (0% BYOK vs their 5%).
Yes. Connect Ollama, vLLM, TGI, or any OpenAI-compatible endpoint. AllRoutes auto-discovers available models, monitors health, and falls back to cloud if your server goes down.
Zero Data Retention by default. Prompts and completions are never logged unless you opt in. Provider-level data policies are enforced. EU data residency available on Team and Enterprise plans.
Email us at [email protected] for any questions. Pro and Team plans include priority support. Enterprise plans include a dedicated Slack channel and 24/7 coverage.
We'd love to hear from you
Have questions? Want a demo? Need a custom plan? Reach out.
Enterprise Sales
Custom plans, on-prem deployment, dedicated SLA, and priority onboarding.
Talk to salesStop paying the AI infrastructure tax.
Join the waitlist. Get $50 free credits at launch. No credit card required.
No credit card. Cancel anytime. Credits never expire.