Gateway

One Gateway for
Every LLM

Route, observe, cache, and orchestrate all your LLM calls through a single control plane. Connect any provider. Ship faster.

No credit card · 5,000 Gateway credits/month · billed on console.phthos.ai

Just change your base URL
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://gateway.phthos.ai/v1",
  apiKey:  "pk_your_api_key",
});

Works with every LLM provider

OpenAIAnthropicGoogle GeminiMistralGroqAzureAWS BedrockOllama

Everything you need to
ship LLM apps

Multi-Provider Gateway

OpenAI-compatible API. Route across OpenAI, Anthropic, Gemini, Groq, and more.

Smart Caching

Exact and semantic response caching to cut latency and provider spend.

RAG & Vector Stores

Ingest documents, search vectors, and ground model answers in your data.

Agent Workflows

Orchestrate tools, memory, and models — usage metered per operation, not per flow.

Gateway plans

Paid plans unlock Gateway features. Usage is metered in Gateway credits. This is not Task AI or Eval billing.

Free

$0/mo

  • 5,000 credits/month
  • 2 LLM providers
  • Basic logging (7-day)
  • 1 RAG & vector store
  • 1 user
Get Started Free

Pro

$50/mo

  • 500,000 credits/month
  • Semantic + exact cache
  • Observability (30-day)
  • Unlimited users
  • Priority support
Start Gateway Pro

Enterprise

$200/mo

  • Custom included credits
  • 3-month retention
  • Priority support & SLA
Start Enterprise

Gateway usage

Credit costs

Debited on your Gateway account only. Cache hits are free.

Model call1
Embedding callStandalone /v1/embeddings1
Vector DB search1
Memory read / write1
Cache readFree
RAG query2
Vector ingest2
Agent workflowPer step

Pay and track usage at console.phthos.ai/billing. Prices in USD.

Ready to simplify your LLM stack?

Start Building — Free