A standalone HTML chat interface for interacting with AI models via API. No installation, no backend, just a single file hosted.
Try it live: ChatAPI
- Zero Dependencies: Just one HTML file. Easy to host and share.
- Fully Configurable: Set your API key, model, endpoint, and system prompt if you want to.
- UI: Dark mode design with animations and responsive layout.
- Markdown Support: AI responses are rendered with full Markdown (code blocks, tables, lists, etc.).
- Local Storage: Remembers your settings between sessions.
- Works with Any API: Compatible with OpenAI, Anthropic, Mistral, Ollama, OpenRouter, Gemini, etc.
- Edit Prompts: You can now edit your own prompt if you did a typo or worded it incorrectly.
- Export Chat: You also can download your AI chat in .json for all your needs incase it's for training a model or saving it for later.
- Chat Storage: When you exit the website or reload the page, it'll still be saved for later.
- Open the App
Visit the hosted page: ChatAPI - Open Settings
Click the Settings button in the top-right. - Enter Your API Details
- API Key: Your secret key (e.g.,
sk-...) - Model: Model name (e.g.,
gpt-3.5-turbo,claude-3-haiku,llama3) - API Endpoint: Full URL (e.g.,
https://api.openai.com/v1/chat/completions) - System Prompt: Set behavior (e.g., "You are a helpful assistant.")
- API Key: Your secret key (e.g.,
- Start Chatting
Type a message and press Send or Enter.
- All processing happens in your browser, no data is sent to any server.
- Your API key is stored only in localStorage and never shared.
- Ideal for private, AI interactions. (Based on what your API service allows.)
Note: Don't expose your API key. This app runs client-side, so your key stays with you.
Found a bug or want a new feature? Feel free to open an issue or pull request and I'll respond as soon as I can.
| Version | Features and Improvements |
|---|---|
| V1.0 | - Initial release allowing users to interact with any AI model using an API key. - Formatted in Markdown for more readability, supporting headers, code blocks, tables, and etc. |
| V1.1 | - Added a Clear Chat button to easily erase conversation history without reloading the page. - Introduced Retry functionality if the AI fails to generate a response. - Enabled Copy Code button for code blocks, allowing easy copying without manual selection. - Improved UI design for a cleaner, more modern appearance. |
| V1.2 | - Added a Stop Generating button to abort ongoing AI responses. - Enhanced error handling for more reliable performance. - Improved local storage to persist chat history across page reloads and browser sessions. - Refactored codebase for better structure and readability, making it easier for developers to customize. |
| V1.3 | - Added the ability to download chat history in .json format for backup or sharing.- Introduced editable prompts, allowing users to correct typos or rephrase their input after sending. |
| V1.4 | - Improved UI consistency with complete styling for all buttons, including the stop button. - Better code organization with cleaner CSS structure and unnecessary comments removed. - Improved input box structure for a more polished and consistent user interface. |