Free AI API

17 models. No signup. No API keys. No credit card.
OpenAI-compatible. Just call it.

Quick start — one curl, that's it

curl https://api.clervo.dev/v1/chat/completions \
  -H 'content-type: application/json' \
  -H 'idempotency-key: my-first-call' \
  -d '{
    "model": "clervo/fast",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Available Free Models

Model Params Latency Best for
clervo/fast 8B 170ms Speed, classification, extraction
clervo/smart 70B 475ms General reasoning, instructions
clervo/deep 550B MoE ~1s Complex analysis, deep reasoning
groq/llama-3.1-8b-instant 8B 170ms Fastest inference
groq/llama-3.3-70b 70B 475ms Best free quality
groq/qwen3.6-27b 27B 350ms Multilingual, reasoning
groq/gpt-oss-120b 120B 570ms Large model, general purpose
groq/gpt-oss-20b 20B ~400ms Balance of speed and size
sambanova/llama-3.3-70b 70B 680ms Alternate Llama 70B route
nvidia/nemotron-ultra-550b 550B MoE ~1s Largest free model
nvidia/nemotron-super-120b 120B 590ms Reasoning, analysis
nvidia/nemotron-nano-30b 30B 500ms Efficient, good quality
nvidia/deepseek-v4-flash ~200B MoE ~3s DeepSeek V4, deep reasoning
nvidia/llama-4-maverick-17b 17B × 128 experts ~430ms Llama 4 MoE, long context
nvidia/gemma-4-31b 31B ~530ms Google Gemma 4, instruction-tuned
nvidia/mistral-small-4-119b 119B ~560ms Mistral Small 4
nvidia/minimax-m3 MiniMax M3 ~1.1s Reasoning

Free Services (also no payment)

Web Search

curl https://api.clervo.dev/v1/search \
  -H 'content-type: application/json' \
  -d '{"query": "latest AI news"}'

Web Scrape

curl https://api.clervo.dev/v1/scrape \
  -H 'content-type: application/json' \
  -d '{"url": "https://example.com"}'

Speech-to-Text (Whisper)

curl https://api.clervo.dev/v1/audio/transcriptions \
  -F [email protected] \
  -F model=whisper-large-v3

Text-to-Speech

curl https://api.clervo.dev/v1/audio/speech \
  -H 'content-type: application/json' \
  -d '{"input": "Hello from Clervo", "voice": "alloy"}' \
  --output speech.mp3

Image Generation

curl https://api.clervo.dev/v1/images/generations \
  -H 'content-type: application/json' \
  -d '{"prompt": "a red prism splitting light", "size": "1024x1024"}'

Crypto Prices

curl 'https://api.clervo.dev/v1/data/prices?ids=bitcoin,ethereum'

Use with any OpenAI SDK

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.clervo.dev/v1",
    api_key="unused"  # No key needed
)

response = client.chat.completions.create(
    model="clervo/fast",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

TypeScript / Node.js

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.clervo.dev/v1',
  apiKey: 'unused',
});

const res = await client.chat.completions.create({
  model: 'clervo/fast',
  messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(res.choices[0].message.content);

FAQ

Is this really free?

Yes. 17 models are genuinely free during beta. No hidden costs, no rate limits (yet), no account needed. We monetize through paid premium models (Claude, GPT) via x402 USDC payments.

What's the catch?

No catch. Free models are powered by Groq, Nvidia NIM, and SambaNova free tiers. We pass through at zero markup. If free tiers get restricted, we'll give advance notice.

Do I need an API key?

No. Send requests directly. Set api_key to any string if your SDK requires it.

What's x402?

x402 is a protocol for AI agents to pay for API calls using USDC stablecoins. Our paid models use it. Free models don't need any payment.

What's the rate limit?

No hard limit during beta. We rely on upstream provider limits (Groq: ~30 req/min on free tier). If you need higher throughput, ask us.