Starting from scratch
Scaffold a full-stack AI app
in one command.
Choose from Next.js, TanStack, or a Node.js agent starter. Every template ships with Gateway integration, streaming, security, and config pre-wired.
npx create-cencori-app my-app
Already have an app
Works with
every AI framework.
First-class SDK for Vercel AI SDK and TanStack. OpenAI-compatible endpoint for every other framework — LangChain, CrewAI, AutoGen, LlamaIndex, and more.
1 import { cencori } from 'cencori';
2 import { streamText } from 'ai';
3
4 const result = streamText({
5 model: cencori('gpt-4o'),
6 messages,
7 });1 OPENAI_BASE_URL=https://api.cencori.com/v1
2 OPENAI_API_KEY=cake_your_key_hereWorks with LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen, Dify, and any OpenAI-compatible SDK in any language.
The platform
Everything you need
to build AI products.
Instead of stitching together providers, vector stores, billing systems, and security middleware — you get one platform with one API.
Multi-Provider Gateway
Route requests to any LLM through a single OpenAI-compatible endpoint. Automatic fallback, model equivalence mapping, and 14+ providers.
AI Security
Prompt injection detection, PII scanning, and content moderation on every request. Zero-config, active by default.
Agent Compute
Serverless code execution for AI agents. Run tools, fetch data, and execute code without managing infrastructure.
Observability
Full request/response logging, P50/P90/P99 latency, token usage, and cost tracking. See everything in real time.
End-User Billing
Meter token usage, set markup pricing, and collect payments via Stripe Connect. Monetize your AI product from day one.
Rate Limits & Budgets
Per-user rate limits and spending caps. Prevent abuse, control costs, and enforce plans without writing billing code.
Monetize your AI
Bill your users
for every token.
Set your margin on every model. Cencori meters usage, enforces rate plans, and handles Stripe payouts. You focus on building — we do the billing.
Buy tokens at cost, sell at your price. Set a 2x, 5x, or 10x markup per model. The math is automatic.
Users pay you directly via Stripe. Cencori handles metering and invoicing. No PCI compliance, no billing infrastructure.
Free tier, pro tier, enterprise tier — define budgets and rate limits per plan. Enforce them without writing code.
Example economics
At 100K requests/day, that's $800/day margin. Cencori handles the metering, invoicing, and collection.
Works with your stack
AI Frameworks
Editors & Platforms