Get Your Free API Keys

Follow these steps to obtain free API keys from various LLM providers

🎯 Why Multiple Providers?

Each provider offers different free tier limits. By registering with multiple services, you maximize your available tokens and ensure high availability through automatic fallback.

⚡ Quick Start

1. Choose providers from the list below
2. Click the link to visit their website
3. Sign up for a free account
4. Generate an API key
5. Add it in the Admin Panel

Groq

30 RPM 6K TPM

Ultra-fast inference with Llama models. Best for real-time applications.

  • Visit console.groq.com
  • Sign up with GitHub or email
  • Go to API Keys section
  • Create a new key
  • Copy and add to Admin panel
Get API Key →

Cerebras

30 RPM 60K TPM

Access to frontier models like Qwen3-Coder 480B with blazing speed.

  • Visit cloud.cerebras.ai
  • Create a free account
  • Navigate to API section
  • Generate your API key
  • Add to Admin panel
Get API Key →

SambaNova

20 RPM 200K TPD

High-performance Llama models with generous daily limits.

  • Go to cloud.sambanova.ai
  • Register for free
  • Access API dashboard
  • Create API credentials
  • Submit to Admin panel
Get API Key →

OpenRouter

20 RPM Free Models

Gateway to multiple free models including DeepSeek, Kimi, Qwen.

  • Visit openrouter.ai
  • Sign in with your account
  • Go to Keys section
  • Create a new API key
  • Add to Admin panel
Get API Key →

GitHub Models

10 RPM GPT-5 Access

Access to cutting-edge models including GPT-5 through GitHub.

  • Visit GitHub Models
  • Sign in with GitHub account
  • Enable model access
  • Generate personal access token
  • Configure in Admin panel
Get API Key →

Mistral AI

2 RPM 500K TPM

European leader with powerful Mistral Large and Codestral models.

  • Go to console.mistral.ai
  • Create free account
  • Navigate to API Keys
  • Generate new key
  • Add to Admin panel
Get API Key →

Cohere

20 RPM 33 RPD

Specialized in enterprise-grade language models with strong reasoning.

Get API Key →

Cloudflare Workers AI

~18-45M tokens Monthly

Run models at the edge with Cloudflare's global network.

  • Go to Cloudflare Dashboard
  • Create account (free)
  • Enable Workers AI
  • Generate API token with Account ID
  • Format: account_id:token in Admin
Get API Key →

HuggingFace Router

~1-3M tokens Varies

Access thousands of open-source models through HF Inference.

  • Visit huggingface.co
  • Create free account
  • Go to Settings → Tokens
  • Create read token
  • Add to Admin panel
Get API Key →

Zhipu AI (Z.ai)

1M TPD GLM Models

Chinese provider with powerful GLM series models.

  • Visit open.bigmodel.cn
  • Register account
  • Navigate to API management
  • Create API key
  • Add to Admin panel
Get API Key →

NVIDIA NIM

Credits-based Enterprise

NVIDIA's optimized inference service with accelerated models.

  • Go to build.nvidia.com
  • Sign up for free credits
  • Select a model
  • Get API key
  • Configure in Admin
Get API Key →

💡 Pro Tips

📊 Track Your Usage

Monitor your token consumption in the Admin panel to avoid hitting limits unexpectedly.

🔄 Rotate Keys

Add multiple keys per provider to distribute load and increase effective rate limits.

🎯 Prioritize Speed

The auto-router prioritizes faster providers. Add Groq and Cerebras for best latency.

🛡️ Security First

All API keys are encrypted with AES-256-GCM before storage. Never share your unified key.

❓ Frequently Asked Questions

Are these APIs really free?

Yes! All listed providers offer free tiers with varying limits. Some require credit card verification but won't charge unless you exceed free limits.

How does the fallback system work?

When you send a request, the router tries the highest-ranked available model. If it's rate-limited or fails, it automatically tries the next provider in the chain.

Can I use my own API key directly?

Yes! You can also call providers directly using their native APIs. Our unified endpoint simplifies this by handling routing and failover automatically.

What happens when I hit a rate limit?

The system automatically switches to another available provider. You'll experience minimal disruption, and the rate-limited provider is temporarily deprioritized.