Sick of 429 errors, 50-message daily caps and free models that die overnight? NanoGPT puts every top roleplay model on one key: pay only for what you use, no subscription, no account or email needed. Set it up as your Janitor AI proxy in 2 minutes: get a key at nano-gpt.com/api, open API Settings → Proxy, paste the config below, save, refresh.

WHY NANOGPT?
• One API key for every major model: DeepSeek, Claude, Gemini, GPT, GLM, Qwen and more. Switch models without new accounts.
• Pay-as-you-go at list price: no markup, no monthly minimum. A few dollars lasts a long time on cheap roleplay models.
• No daily message caps like free tiers. If a model's provider is busy, switch models with the same key.
• Private by default: no account or email needed. Grab a key, add funds, chat.
• Card or crypto top-ups: crypto gets a 3% bonus.

STEP 1: GET YOUR API KEY
1. Go to nano-gpt.com/api and copy your API key.
2. Add a small balance (or use a NanoGPT subscription).
Never share your key or post it in comments.

STEP 2: OPEN JANITOR'S PROXY SETTINGS
1. Open any character chat.
2. Click "using janitor" (or the menu in the top right) → API Settings.
3. Select Proxy as the provider.
4. Under Proxy Configurations, click + Add (or Import).
The character you're chatting with must allow proxies.

STEP 3: PASTE THIS CONFIG

Import option (replace the key with yours first):
{
"name": "NanoGPT",
"apiUrl": "https://api.nano-gpt.com/api/v1/chat/completions",
"apiKey": "YOUR_NANOGPT_API_KEY",
"model": "auto-model-basic"
}

Manual option:
• Configuration Name: NanoGPT
• Proxy URL: https://api.nano-gpt.com/api/v1/chat/completions
• API Key: your NanoGPT key
• Model Name: auto-model-basic
Keep the full URL, including /chat/completions.

STEP 4: SAVE, ACTIVATE, REFRESH
Click Apply, make sure the NanoGPT config is marked Active, then refresh the page. Send a short test message.

STEP 5: PICK YOUR MODEL
auto-model-basic picks a model for you automatically. To lock in one model, copy its exact ID from nano-gpt.com/pricing and paste it into Model Name.

Want the model's thinking shown in chat? Duplicate the config and change the Proxy URL to:
https://api.nano-gpt.com/api/v1thinking/chat/completions
Tip: make one config per model and name them clearly so you can switch in one click.

RECOMMENDED SETTINGS
• Temperature: 0.8–1.0 for roleplay
• Max tokens: 0 for reasoning models; 800–1000 for others
• Context size: 32k (higher = better memory, but costs more per message)
Every setting explained: Janitor AI API Settings Guide (2026)

TROUBLESHOOTING

"No response from bot" / proxy error: you picked a paid model and your balance is empty, or the model isn't included in your subscription. Top up or switch models.
401: API key is wrong. Re-copy it with no extra spaces.
404: the model ID is outdated or misspelled. Copy it fresh from nano-gpt.com/pricing, and check you didn't add /chat/completions twice.
429 / 503: that model's upstream provider is busy. Wait a minute or switch models; your key works with all of them.
"Failed to fetch": refresh Janitor, make sure the NanoGPT config is Active, and turn off any VPN or extension blocking requests.
No proxy option: the bot's creator disabled proxies. Try a different character.

MORE HELP
Quick 5-step setup: Janitor AI API Tutorial
API settings explained: Janitor AI API Settings Guide (2026)
Official guide: docs.nano-gpt.com/integrations/janitorai
Models and prices: nano-gpt.com/pricing
Community: NanoGPT Discord

This guide gets updated when anything changes.