Free LLM API — Unified Multi-Provider AI Gateway

from openai import OpenAI

# Point the SDK at the LLM router instead of OpenAI
client = OpenAI(
    base_url="https://agent-gateway-kappa.vercel.app/v1/agent-llm",
    api_key="YOUR_GATEWAY_KEY"  # your Agent Gateway key
)

# Use any supported model — even non-OpenAI ones
response = client.chat.completions.create(
    model="gpt-4o-mini",
    messages=[
        {"role": "user", "content": "Write a haiku about APIs"}
    ],
    extra_headers={"X-Provider-Key": "sk-your-openai-key"}
)

print(response.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://agent-gateway-kappa.vercel.app/v1/agent-llm",
  apiKey: "YOUR_GATEWAY_KEY",
  defaultHeaders: { "X-Provider-Key": "sk-your-openai-key" }
});

const completion = await client.chat.completions.create({
  model: "gpt-4o-mini",
  messages: [{ role: "user", content: "Write a haiku about APIs" }]
});

console.log(completion.choices[0].message.content);

curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-20250514",
    "messages": [
      {"role": "system", "content": "You are a concise technical writer."},
      {"role": "user", "content": "Compare REST and GraphQL in 3 bullets."}
    ],
    "provider_key": "sk-ant-your-anthropic-key",
    "max_tokens": 500
  }'

curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-2.0-flash",
    "messages": [
      {"role": "user", "content": "What is the capital of France?"}
    ],
    "provider_key": "your-google-api-key"
  }'

curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama-3.3-70b-versatile",
    "messages": [
      {"role": "user", "content": "Explain recursion with a Python example."}
    ],
    "provider_key": "gsk_your-groq-key"
  }'

curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-reasoner",
    "messages": [
      {"role": "user", "content": "Solve: What is the derivative of x^3 * sin(x)?"}
    ],
    "provider_key": "sk-your-deepseek-key"
  }'

# This request will be cached
curl -X POST https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"What is 2+2?"}],"provider_key":"sk-..."}'

# Same request within 5 min — instant response from cache
# Response includes: "cached": true

import requests

BASE = "https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/chat"

def llm_call(model, messages, provider_key):
    resp = requests.post(BASE, json={
        "model": model,
        "messages": messages,
        "provider_key": provider_key
    })
    return resp.json()["choices"][0]["message"]["content"]

user_question = "Explain the difference between TCP and UDP"

# Step 1: Classify complexity with a fast/cheap model
complexity = llm_call(
    model="llama-3.1-8b-instant",
    messages=[{"role": "user", "content":
        f"Rate this question's complexity as 'simple' or 'complex': {user_question}"}],
    provider_key="gsk_your-groq-key"
)

# Step 2: Route to appropriate model
if "complex" in complexity.lower():
    answer = llm_call(
        model="gpt-4o",
        messages=[{"role": "user", "content": user_question}],
        provider_key="sk-your-openai-key"
    )
else:
    answer = llm_call(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": user_question}],
        provider_key="sk-your-openai-key"
    )

print(f"Complexity: {complexity}")
print(f"Answer: {answer}")

# List all 24+ models with provider info
curl https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/models | jq .

# List providers and their model counts
curl https://agent-gateway-kappa.vercel.app/v1/agent-llm/api/providers | jq .

{
  "models": [
    {"model": "gpt-4o", "provider": "openai", "providerName": "OpenAI"},
    {"model": "gpt-4o-mini", "provider": "openai", "providerName": "OpenAI"},
    {"model": "claude-opus-4-20250514", "provider": "anthropic", "providerName": "Anthropic"},
    {"model": "gemini-2.0-flash", "provider": "google", "providerName": "Google (Gemini)"},
    {"model": "llama-3.3-70b-versatile", "provider": "groq", "providerName": "Groq"},
    {"model": "deepseek-chat", "provider": "deepseek", "providerName": "DeepSeek"}
  ],
  "total": 24
}

# Missing provider key
{"error": "Provider API key required. Pass as \"provider_key\" in body or set via POST /api/keys/provider"}

# Unknown model
{"error": "Unknown model: gpt-5. Use GET /api/models to see available models."}

# Upstream provider error
{"error": "OpenAI API error 429: Rate limit exceeded"}