New60+ models live, metered per token

Every major AI model. One API key.

ModelPort routes your requests to the best models on the market and bills you only for the tokens you use — no subscriptions, no per-provider accounts.

Start building View pricing

quickstart.py

from openai import OpenAI

client = OpenAI(
    base_url="https://modelport.tech/v1",
    api_key="mp_live_...",
)

client.chat.completions.create(
    model="claude-sonnet-4",
    messages=[{"role": "user", "content": "Hello"}],
)

OpenAI Anthropic Google DeepSeek Meta Mistral xAI Qwen MiniMax Zhipu OpenAI Anthropic Google DeepSeek Meta Mistral xAI Qwen MiniMax Zhipu

Built for developers who ship

The infrastructure between you and every major model provider, reduced to one endpoint and one balance.

One key, every model

GPT, Claude, Gemini, DeepSeek and dozens more behind a single OpenAI-compatible endpoint. Switch models by changing one string.

Pay per token

No subscriptions, no seats, no minimums. Top up once and spend only on the tokens you actually consume.

Zero-retention mode

Flip privacy mode on any API key and nothing is stored but the token counter required for billing.

Real-time metering

Balance and usage update the instant a request completes. Every token accounted for, down to fractions of a cent.

No rate surprises

Per-model prices are listed publicly and billed exactly. What you see on the pricing page is what hits your balance.

Production-grade auth

Hashed keys, revocable in one click, per-key last-used tracking. Rotate credentials without downtime.

Up and running in minutes

Create an account

Top up your balance

Card or crypto, starting at $5. Funds credit instantly and never expire.

Point your app at us

Change one base URL in your existing OpenAI SDK code. That's the whole migration.

Stop juggling provider accounts

One signup, one balance, one key. Your first $5 of credit goes further than you think.

Create free account