Connect Your API Key
Set up your own AI provider in under 2 minutes
Unlimited Usage
No monthly credit limits
Your Data, Your Key
Direct to AI provider
12 Models
OpenAI, Claude, Gemini, OpenRouter, or your own endpoint
OpenAI
GPT-5.4, GPT-4o, and more
- 1
Go to OpenAI API Keys
platform.openai.com/api-keys - 2
Click “Create new secret key”
Give it a name like “KeyShift AI”
- 3
Copy the key (starts with
sk-)You won't be able to see it again
- 4
In KeyShift AI: Settings → AI Provider → Paste key → Save
Click “Test Connection” to verify
Recommended model: GPT-5.4 Mini — best balance of speed and quality. ~$0.001 per request.
Anthropic (Claude)
Claude Opus 4.6, Sonnet, Haiku
- 1
Go to Anthropic Console
console.anthropic.com/settings/keys - 2
Click “Create Key”
Name it “KeyShift AI”
- 3
Copy the key (starts with
sk-ant-) - 4
In KeyShift AI: Settings → AI Provider → Select “Anthropic” → Paste key → Save
Recommended model: Claude Haiku 4.5 — fastest and cheapest. ~$0.001 per request.
Google Gemini
Gemini 2.5 Pro, Flash, Flash-Lite
- 1
Go to Google AI Studio
aistudio.google.com/apikey - 2
Click “Create API Key”
Select or create a Google Cloud project when prompted
- 3
Copy the API key
Starts with
AIza - 4
In KeyShift AI: Settings → AI Provider → Select “Google (Gemini)” → Paste key → Save
Recommended model: Gemini 2.0 Flash — fast, cheap, and great quality. Free tier available with generous limits.
OpenRouter
Hundreds of models from one key
- 1
Go to OpenRouter Keys
openrouter.ai/keys - 2
Click “Create Key” and add credits
You pay OpenRouter directly, usually cents per day
- 3
Copy the API key
Starts with
sk-or- - 4
In KeyShift AI: Settings → AI Provider → Select “OpenRouter” → Paste key → Save
The model list loads live from OpenRouter, so you always see what's available
Why OpenRouter: one key gives you GPT, Claude, Gemini, Llama, DeepSeek and many more, including free models. Switch models anytime without creating new accounts.
Custom endpoint
Your own server, a local model, or any OpenAI-compatible API
- 1
Settings → AI Provider → Select “Custom endpoint (OpenAI-compatible)”
- 2
Paste your endpoint URL
Either form works:
https://your-server.com/v1or the full/chat/completionspath - 3
Enter the model id your server serves
Not sure? Open
YOUR_URL/modelsin a browser to list them - 4
Add an API key if your endpoint needs one, then Test Connection
Leave the key empty for self-hosted servers with no auth
Works with: Ollama (http://localhost:11434/v1), LM Studio, vLLM, LiteLLM, Azure OpenAI, or your own proxy in front of any provider. The endpoint has to speak the OpenAI chat/completions format.
What does it actually cost?
Estimates based on GPT-4o-mini pricing. Costs vary by model and request length.
Troubleshooting
“Invalid API key”
Make sure you copied the full key. For OpenAI, it starts with sk-. For Anthropic, it starts with sk-ant-. Check that you selected the correct provider in the dropdown.
“Connection failed”
Check your internet connection. If you're behind a corporate firewall or VPN, the provider's API endpoint may be blocked. Try disabling your VPN temporarily.
“Rate limit exceeded”
You've hit the provider's rate limit. Wait a moment and try again. New OpenAI accounts have lower limits — they increase automatically after a few days of usage and billing.
AI Provider section not visible
The API key settings only appear for Lifetime and AppSumo Tier 3 users. If you're on a different plan, upgrade to unlock BYOK.