The brain behind

One API. Every AI model. Smart routing, real-time cost tracking, and complete control over your agent's AI spend.

Start FreeView Documentation

500 free credits · No credit card required

terminal
|

How cortex works

1

Connect

Replace your provider's base URL with cortex. One line change.

-baseURL: "https://api.openai.com/v1"
+baseURL: "https://cortex.zoneout.io/api/v1"
2

Route

cortex picks the best model for each request. Optimize for quality, speed, or cost.

3

Control

Set budgets, track spend in real-time, get alerts before limits hit. Never overspend.

Works with every agent

cortex is the infrastructure layer. Your agent keeps full control — we just make it smarter and cheaper.

O

OpenClaw

The most popular open-source AI agent. 160K+ GitHub stars.

Compatible
M

Manus AI

Autonomous execution agent acquired by Meta. Deploys apps, browses web.

Compatible
N

NanoBot

Lightweight agent in 4K lines of Python. 8 messaging channels.

Compatible
C

Claude Code

Anthropic's CLI coding agent. Multi-file edits, PR workflows.

Compatible
C

Cursor

AI-native code editor with inline completions and background agents.

Compatible
S

SuperAGI

Multi-agent framework for orchestrating autonomous AI teams.

Compatible
N

NanoClaw

Security-first agent with container isolation and sandboxed execution.

Compatible
n

n8n

Open-source workflow automation. Visual builder with 400+ integrations.

Compatible
L

Lindy.ai

Consumer AI assistant via iMessage. 400K+ users.

Compatible
C

CrewAI

Framework for orchestrating role-playing AI agents.

Compatible

Don't see your agent? cortex works with any tool that calls an OpenAI-compatible API.

40+ models, 6 providers, one API

AnthropicOpenAIGoogle AIxAIReplicateElevenLabs

Light

Fast & cheap
  • Haiku 4.5
  • GPT-4o-mini
  • Gemini 2.0 Flash
  • Grok 3 Mini

Standard

Balanced
  • Sonnet 4.5
  • GPT-5
  • GPT-4.1
  • Gemini 2.5 Pro
  • Grok 3

Premium

Maximum intelligence
  • Opus 4.6
  • o3
  • GPT-5.3
  • Gemini 3.1

Plus image generation (FLUX, DALL-E), speech-to-text (Whisper), text-to-speech (ElevenLabs)

Glass box, not black box

See every sub-task, which model runs it, and what it costs — in real-time

Replay any request: full input/output, latency, token count

Set per-workflow budgets: cortex stops before you overspend

Export any workflow as reproducible API calls

Other platforms decide everything for you. cortex shows you everything and lets you decide.

✓
ResearchGemini
1.2 credits4.1s
✓
AnalyzeSonnet
1.8 credits5.6s
✓
WriteGPT-5
1.2 credits2.6s
Total4.2 credits · 12.3s · 3 models

Why cortex?

FeaturecortexOpenRouterDirect API
Smart routing
Cost tracking
Per-agent budgets
Spend alerts
Model fallback
Dashboard analytics
OpenAI compatible
PriceFrom $0/moPay per tokenPer provider billing

Simple, transparent pricing

Free

$0/mo

500 credits/month

  • Light models
  • 10 req/min
  • 1 API key
  • Community support
Get Started
Most popular

Spark

$29/mo

5,000 credits/month

  • Light + Standard models
  • 30 req/min
  • 2 API keys
  • Email support
Subscribe

Pro

$99/mo

25,000 credits/month

  • All models
  • 100 req/min
  • 5 API keys
  • Priority support
  • Advanced analytics
Subscribe

Ultra

$349/mo

100,000 credits/month

  • All models + priority
  • 500 req/min
  • Unlimited API keys
  • Dedicated support
  • Custom routing rules
Subscribe

Need more? Buy credit packs starting at $12 for 1,000 credits. Purchased credits never expire.

Trusted by builders

“

cortex cut our agent's API costs by 40% with smart routing alone.

A

Alex

AI Engineer

“

Finally I can see exactly where my credits go. The glass box approach is a game changer.

S

Sarah

Indie Developer

“

Switched from managing 4 provider accounts to just cortex. Setup took 5 minutes.

M

Marcus

Startup CTO

FAQ

Ready to give your agents a brain?

Start free. Scale when you need to.

Create Free Account

500 credits · No credit card · Setup in 60 seconds