🌐 NextChat (Web/Desktop) Integration Guide
NextChat (formerly ChatGPT Next Web) is a lightweight, cross‑platform open‑source chat client. It offers Windows/Mac desktop apps and can also be deployed as a web app. If you prefer a minimal interface, it’s a solid choice.
📥 Step 1: Access or download
- Web: you can use any publicly hosted NextChat instance (it is front‑end only; your key is stored locally in the browser).
- Desktop releases: NextChat GitHub Releases
⚙️ Step 2: Configure endpoint & API key
- Open NextChat and click Settings (gear icon) in the bottom-left.
- Scroll down to Model Provider and choose
OpenAI. - Enable Custom Endpoint.
- Fill in:
Endpoint URL: our base request domain:
https://api.olddog.shop(⚠️ Note: NextChat usually appends paths automatically, so you typically do not need
/v1here. If it doesn’t work, try adding/v1.)API Key: your token generated in the Old Dog AI Console (starts with
sk-).
🧩 Step 3: Add custom models (important)
NextChat’s default dropdown usually only includes older OpenAI models. To use claude-sonnet-4-6 or gpt-5.2, add them manually:
- Continue scrolling in settings to Custom Models.
- Enter the model names you want (refer to the recommended names in Models & Pricing).
- Rules: separate multiple models with an English comma
,. To hide built‑in models, prefix with-all.- Example (copy/paste):
-all,+claude-opus-4-6,+claude-sonnet-4-6,+gpt-5.2,+o3-pro,+gemini-3.2-pro
- Example (copy/paste):
- NextChat saves automatically.
💬 Step 4: Select a model and start chatting
- Return to the main chat view.
- Click the model name at the top (center or top-right).
- You should see your added models like
claude-sonnet-4-6. - Select one and start.
💡 Troubleshooting
- Spinning then “Network Error”: often caused by proxy conflicts in certain network environments or an incorrect endpoint. Double‑check the endpoint spelling.
- Invalid API Key: your key is incorrect or contains extra spaces/newlines. Copy it again from the console.
