Janitor AI API tutorial: connect NanoGPT's API to Janitor AI in 5 steps. Open a chat → API Settings → Proxy → paste your NanoGPT API key, proxy URL and model → save and refresh. One API key gives you DeepSeek, Claude, Gemini, GPT, GLM and more, a simple OpenRouter alternative with no daily caps.

1. GET YOUR NANOGPT API KEY
Go to nano-gpt.com/api and copy your API key. No account or email needed.
Add a small balance. Pay-as-you-go at list price, no markup, no daily message cap.

2. OPEN API SETTINGS IN JANITOR AI
Open any chat. Then open API Settings:
Menu → API Settings, or click "using janitor" at the top of the chat.

3. PICK PROXY
Under Provider, choose Proxy. Then click + Add under Proxy Configurations.

4. SET THE FOLLOWING
• Configuration Name: NanoGPT
• Model Name: auto-model-basic (or any model ID from nano-gpt.com/pricing)
• Proxy URL: https://api.nano-gpt.com/api/v1/chat/completions
• API Key: your NanoGPT API key

5. SAVE AND REFRESH
Click Apply, make sure the NanoGPT config is Active, then refresh the page (required). Send a test message. Done.

SWITCHING FROM OPENROUTER?
Same setup, two changes: swap the Proxy URL to https://api.nano-gpt.com/api/v1/chat/completions and paste your NanoGPT API key instead. No 50-message daily cap, same popular models, one key, pay-as-you-go.
Model IDs can differ slightly between providers, so copy the exact ID from nano-gpt.com/pricing or just use auto-model-basic.

QUICK TIPS
• Switch models: make one config per model ID and name them clearly. Same API key works for all of them.
• Show the model's thinking in chat: use https://api.nano-gpt.com/api/v1thinking/chat/completions as the Proxy URL.
• Starter settings: Temperature 0.9, Max tokens 0 for reasoning models (800–1000 for others), Context 32k.
• Proxies blocked? Some bot creators disable proxies. Use a character marked "proxy allowed".

COMMON API ERRORS
401: wrong API key. Re-copy it with no extra spaces.
404: wrong model ID or URL. Copy the ID fresh from the pricing page, and don't add /chat/completions twice.
429 / 503: the model's provider is busy. Wait a minute or switch models.
"No response from bot": balance is empty. Top up or switch to a model in your subscription.
"Failed to fetch": refresh Janitor and turn off any VPN or extension blocking requests.

MORE
Full proxy guide: Best Proxy for Janitor AI (2026)
Every API setting explained: Janitor AI API Settings Guide (2026)
Official docs: docs.nano-gpt.com/integrations/janitorai
Community: NanoGPT Discord