NextChat
NextChat (formerly ChatGPT-Next-Web) is the most popular open-source cross-platform AI client on GitHub (75k+ stars). It supports one-click Vercel/Docker deployment or running directly as a cross-platform desktop application. Featuring a lightweight and agile UI with a rich built-in library of professional masks (Prompt presets), it connects seamlessly to the OpenRely API to quickly start conversations with top global large models.
🔑 Prerequisites
sk-...🚀 Steps
Launch the Client or Open the Web Page
You can use NextChat in two ways:
Go to NextChat GitHub Releases to download the latest installer for your operating system (supports macOS, Windows, and Linux), and launch it after installation.
If you have privately deployed NextChat via Vercel or Docker, simply open the deployed site domain in your browser.
Open API Settings
- Click the Settings Gear Icon ⚙️ in the lower-left corner of the NextChat interface;
- Expand the Model Provider section;
- Toggle on Custom Endpoint.
Configure OpenRely Service Parameters
Fill in the parameters according to the table below:
| Configuration | Recommended Value | Description |
|---|---|---|
| API Endpoint | https://api.openrely.ai | Note: Do NOT append /v1 at the end; NextChat appends it automatically under the hood |
| API Key | sk-... | Your dedicated access token created in the OpenRely console |
| Models | +claude,+gpt,+gemini | Prefix the model name with + to append target OpenRely models to the selection menu |
[!TIP] If you want to clear the default official model list and keep only your specified models, enter
-all,+gpt,+claude(check the console or pricing page for specific model IDs).
Verify and Start Chatting
Return to the main chat window:
- Click the model selection dropdown in the top header and switch to the OpenRely model you just added;
- Send a test message (e.g., "Please write a quicksort algorithm in Python");
- If a smooth typewriter-style streaming response is displayed, the integration is successful!
💡 Core Features
| Feature | Description |
|---|---|
| Out-of-the-Box Mask Library | Built-in 100+ ready-to-use Prompt mask personas covering programming, translation, academic editing, and brainstorming |
| Client-Side Local Privacy Storage | Chat history is fully encrypted and stored locally in your browser or client, never uploaded to any third-party storage server |
| Multimodal & Vision | Supports uploading images and asking questions with vision-capable multimodal models (such as GPT-4o, Claude 3.5 Sonnet) |
| Comprehensive Message Export | One-click export of high-quality conversations as long screenshots, Markdown documents, or JSON data backups |
❓ Frequently Asked Questions
Encountering other API errors or billing questions? Please refer to the complete FAQ and Troubleshooting Guide.
Was this page helpful?
Cherry Studio
Cherry Studio is a cross-platform desktop AI assistant client built for professional users. It features 30+ built-in industry agent templates, supports multi-model concurrent comparison, local knowledge bases, document parsing, and long-context reasoning. Connect to the OpenRely API to schedule top global LLMs in one stop.
Chatbox
Chatbox is an established cross-platform desktop AI client favored by professionals and developers. Without requiring any complex development or maintenance environment, simply double-click the installer to launch quickly on Windows, macOS, Linux, iOS, and Android. Connect to the OpenRely API to orchestrate top global foundation models in one place.