Free Groq API Key (Nov 2026): 30 RPM, Ultra-Fast LPU & No Credit Card
Looking for the fastest free LLM inference on the planet? Groq powers models using custom LPU (Language Processing Unit) silicon, delivering speeds up to 280 to 1,200 tokens/second. Every developer receives a generous free tier of up to 30 RPM and 14,400 requests/day with zero credit card required.
How to Get Your Free Groq API Key in 60 Seconds
- Go to Groq Console: Navigate to console.groq.com/keys.
- Sign In: Log in with your GitHub or Google account (or create an account with any email address). No credit card or billing details are ever requested.
- Create Key: Click the Create API Key button, give it a descriptive name (e.g.,
cursor-devoragent-key). - Copy Token: Copy your key immediately (starts with
gsk_). Save it in your local.envfile.
🌍 Regional Availability & Geo-blocking Notice
Groq is available across 180+ countries (including the US, EU, UK, India, and Japan). However, due to US export compliance, direct requests from Mainland China, Hong Kong, and OFAC-sanctioned regions are blocked at the Cloudflare edge layer with HTTP 403 Forbidden.
Need an unrestricted global alternative? Use Cloudflare Workers AI (runs across 300+ edge locations worldwide without regional blocking) or OpenRouter.
How to Call Groq API via Python (OpenAI SDK)
Because Groq is 100% OpenAI-compatible, you only need to change the base_url and pass your Groq key:
import os
from openai import OpenAI
# Initialize client with Groq endpoint
client = OpenAI(
base_url="https://api.groq.com/openai/v1",
api_key=os.getenv("GROQ_API_KEY", "gsk_YOUR_KEY_HERE")
)
# Stream completion at lightning speed
response = client.chat.completions.create(
model="llama-3.3-70b-versatile",
messages=[
{"role": "system", "content": "You are an expert software engineer."},
{"role": "user", "content": "Write a clean binary search in TypeScript."}
],
temperature=0.6,
stream=True
)
for chunk in response:
content = chunk.choices[0].delta.content or ""
print(content, end="", flush=True)Connecting Groq to Cursor, OpenCode, Kilo Code, and Aider
1. Cursor IDE Configuration
- Open Cursor → Settings → Models.
- Turn Override OpenAI Base URL to
ON. - Set Base URL:
https://api.groq.com/openai/v1 - Under OpenAI API Key, paste your
gsk_...key. - Add Model ID:
llama-3.3-70b-versatile.
2. OpenCode & Kilo Code Configuration
In your tool configuration (e.g., ~/.config/opencode/config.json or Kilo Code environment):
{
"provider": "openai",
"base_url": "https://api.groq.com/openai/v1",
"api_key": "gsk_YOUR_KEY_HERE",
"model": "llama-3.3-70b-versatile"
}3. Aider CLI & Claude Code CLI Configuration
# For Aider CLI
export OPENAI_API_BASE="https://api.groq.com/openai/v1"
export OPENAI_API_KEY="gsk_YOUR_KEY_HERE"
aider --model openai/llama-3.3-70b-versatile
# For Claude Code CLI (via LiteLLM proxy)
export ANTHROPIC_BASE_URL="http://localhost:4000"Frequently Asked Questions (FAQ)
How do I get a free Groq API key in 2026?
Go to console.groq.com/keys, sign in with your email or Google account, click "Create API Key", name your key, and copy the generated token (starts with gsk_). It takes less than 60 seconds with zero credit card or phone verification required.
Is Groq API completely free, and does it require a credit card?
Yes, Groq provides a generous permanent free tier with 0 credit card and 0 payment info required. Free tier quotas include up to 30 Requests Per Minute (RPM) and 14,400 Requests Per Day (RPD) depending on the model.
Why is Groq so much faster than other LLM cloud providers?
Groq runs on custom LPU (Language Processing Unit) tensor silicon instead of traditional Nvidia GPUs. This architecture delivers deterministic low-latency token streaming, reaching between 250 to 1,200 tokens/second on supported models.
Can I use Groq as a drop-in replacement for OpenAI in Cursor or Python?
Yes. Groq provides an official OpenAI-compatible endpoint at https://api.groq.com/openai/v1. You can drop it directly into Cursor IDE (Settings -> Models -> Override OpenAI Base URL), Claude Code (via ANTHROPIC_BASE_URL gateway), or any standard openai Python / Node.js library.
How do I resolve Groq 429 RateLimitError (TPM / RPM limits)?
When you hit Groq's rate limits (HTTP 429), implement exponential backoff with jitter, or fail over to alternative free providers like Cloudflare Workers AI or NVIDIA NIM. Check the rate limits section in this guide for tested retry logic.
Does Groq have geographic or regional IP restrictions?
Yes. Groq complies with US export regulations and Cloudflare edge security filters. While Groq is accessible across 180+ countries (including North America, the EU, UK, India, and Japan), access from Mainland China, Hong Kong, and OFAC-sanctioned regions will return HTTP 403 Forbidden errors. If you are in a restricted region, use Cloudflare Workers AI or OpenRouter as a globally distributed, unrestricted free alternative.