LinGPT
A ChatGPT (GPT-3.5 / GPT-4) Webpage with Just a Single HTML File
ChatGPT chat page with only one html file
Demo • Download • Download
Give a Star ? if you like it.
LinGPT
- Introduction: GPT with only one
.html file, lightweight, convenient, zero threshold - Online: Static web pages, small size, hosted on Github/Gitee very smooth
- Download: Download to the computer, open the html file and use it, as simple as opening a picture
- Upload: Can be uploaded to hosting platform/host/server, supporting all free static web hosting platforms
- Interface: Supports OpenAI official API and third-party platform API (already adapted to common proxy platforms), and supports magic-free
- Model: Supports all the latest GPT-3.5 and GPT-4 models
- Key: Supports OpenAI official key and third-party platform dedicated key, and supports polling
- Data: Only stored in the browser local cache, supports export, import and deletion
Online experience
Line Github: https://lin2025.github.io/gpt3.5/
??Line Gitee: https://lin2025.gitee.io/gpt3.5/
??Multiple download methods (including network disk): https://lin2025.gitee.io/#gpt3.5-code
Function
GPT responses are generated by one-time return, non-streaming response, and the intercommunication experience will be reduced, but it also has advantages - it will not interrupt, and the content will be fast when it is too much.
New features : Memory mode (context mode)
New features : Supports dedicated keys for third-party platforms, and adapts to some third-party proxy platforms (api2d, next-web, openai-sb, aiproxy, openaimax, aigc2d, api2gpt, chatai, closeai, ohmygpt, aiayw, czl...)
New features : Adapt to the latest models, supporting all GPT-3.5 and GPT-4 models
New feature : Support automatic requestion (when sending fails, intelligently determine whether to try again)
New features : Polling Key, fully automatic intelligent rotation API Key, support batch addition
New features : Page zoom (mobile terminal)
New features : multi-dialog management, fully automatic saving, support for creating copy, multi-tab data synchronization, exportable (backup/migration/sharing), importable (restore conversations)
New features : Supports browser local storage, supports data synchronization, import, export, deletion, statistics and other operations
Smart max_tokens (full automatic adjustment of max_tokens)
API Key, API balance/validity period query, custom interface URL, custom API request parameters (model, max_tokens, temperature, top_p, present_penalty, frequency_penalty)
Support Markdown (code blocks, highlight codes, display network pictures, display tables, and article typesetting)
System prompt words, Token usage statistics, one-click copying, automatic analysis of OpenAI error codes, and functional help
Clear, undo, reask, and support shortcut keys
The language supports simplified Chinese and English, supports multiple ways to set avatars (can upload avatars), supports style settings (font size, WeChat style time, message time, and distinguish multiple rounds of conversations)
Single-wheel chat records, all chat records and operation records can be exported. The format is Markdown, and it supports exporting .md and .txt
Features
Use notepad or text editing.app to modify the code. The tutorial is written in the code without understanding the technology
Most personalized settings can be completed directly in the interface from v6.07
- No threshold can be used to DIY for everyone. Xiaobai can also use the free static hosting platform to easily build a website
- Supports invisible system prompt words to easily customize AI tools for different purposes in one minute
- Many important functions are under study for daily mild and moderate use scenarios...
renew
- August 28 New memory mode (context mode).
- August 21 1) Upgrade to support GPT-4; 2) Support third-party API Key and adapt to third-party proxy platforms.
- July 29 1) Upgrade to support the latest model, support 16K model; 2) Optimize data management functions: can display conversation size, support erase logs, improve prompts, etc.
- July 23 1) Added the polling key function, intelligent rotation API Key; 2) Added the automatic requestion function to intelligently analyze to determine whether to resend the request; 3) Enhanced error code recognition, fix the problem of modifying the default language in the code, fix the problem of partial input methods that cannot be wrapped in the mobile terminal, fix compatibility issues such as font blur in Windows, fix chat box bugs and optimize styles, fix other style problems, fix other bugs; optimize the comprehensive experience.
- July 9 Optimization of mobile compatibility; fixing style issues; adding mobile page zoom function.
- July 05 v7.05 1) Added local storage and data management functions; 2) Added multiple conversations, automatic saving, copy creation, data synchronization, export sharing, import and restore conversations and other functions; 3) Fixed compatibility issues in Windows system (scroll bars and drop-down menu style errors); 4) Optimize UI.
- June 15th Added avatar upload function. Smart max_tokens upgrades the new algorithm and adds the function of automatically resend after calculation errors. Adjust css compatibility.
- On June 12 , the Smart max_tokens function was added, and
max_tokens was fully adjusted automatically. Fixed the problem: Prompt words allow saving blanks (it was accidentally restricted before). - June 9th, an OpenAI error code was added; fix some problems: problems caused by the use of relative paths of the avatar, problems of bubble frame collapse during loading, and problems that some browsers do not support fonts below 12px.
- On June 7 , the almost brand new version of v6.07 has a relatively large change, with functions upgraded, experience upgraded, and more applicable scenarios, and it is also very stable and no errors are reported.
Updates before June 7
- May 17th The changes are not big, small updates: add
<!DOCTYPE html> to adjust irregular styles; patch the Markdown plug-in Marked.js will filter out some chat content bugs; add small arrows to the bubble box. - May 14 Updated content: Optimize loading speed & stability; fix strobes; optimize layout & experience & increase space; reduce when the chat box loses focus; modify the logic of undoing & repeated questions; correct the bug where the failed message will be written to the context; undoing & repeated questions can be updated in real time; buttons use SVG icon; use random greetings to verify API-key; code optimization & correct naming & increase readability...
- May 9th One-click copy clipboard update content: supports one-click copy, use SVG to replace invalid logos, optimize code, and fix small bugs.
- On May 6, Markdown Highlight is an important update, supporting Markdown + Highlight code highlighting. Debugged for a long time, and the effect is pretty good. In the near future, some practical functions will be added, and the application scenarios can increase from mild to moderate.
- April 28 Add English comments.
- April 27 New function - Query the balance of OpenAI account After detecting API-Key, click [Tokens] to query the balance of API account, use the official OpenAI interface to return the data including the registered name, whether to bind a card, total amount, balance, validity period, etc. The code is adjusted and implemented based on the projects of the following two big shots: @ClarenceDan's openai-billing and @herobrine19's openai-billing. It has been commented in detail, safe and transparent, and has no risk of data leakage.
- April 26th Add dynamic effects to the status of the two buttons "Detection" and "Send"; API-Key clear code will no longer be displayed on the page; some minor errors will be corrected.
- April 23rd Important fixes bug fixes fixes that the web page error caused by IPs in many countries (Vue and Axios cannot be loaded). The Unpkg CDN line has been added. When BootCDN is hung up, it will automatically switch to Unpkg.
- April 16 Modify the shortcut keys on the PC side, change them to Enter to send, and support multiple line breaking methods. It is adapted to the Chinese input habit. In the Chinese input state, the carriage return is confirmed but not sent , and the carriage return will be sent only after the carriage return is received.
- ...
LinGPT
- LinGPT is a lightweight single-page chat application that can be deployed statically, with no need for a server.
- Can be uploaded to any static website hosting platform like GitHub and provide a very smooth browser experience.
- Download and open the index.html file on your computer, input API Key to get started.
- Supports OpenAI API and Third-party APIs (reverse proxy).
- Supports the latest GPT-3.5 and GPT-4 models.
- Supports both OpenAI API Key and Third-party proxy keys. Supports fully automatic intelligent rotation of API keys.
- User data is stored only in the browser's local cache, supporting export, import, and deletion.
Demo
https://lin2025.github.io/gpt3.5/
Features
This code does not use streaming response, GPT will generate the complete response text at once.
- New Features : Contextual mode.
- New Features : Support for third-party API keys, optimized for compatibility with third-party proxy platforms.
- New Features : Support for all models of GPT-3.5 and GPT-4.
- New Features : Supports Automatic Retry (intelligently determines whether to retry when a request fails).
- New Features : Supports Auto Key Rotation, intelligently rotates API keys, supports bulk addition.
- New Features : Page Scaling (For Mobile).
- New Features : Multiple chats management, automatic saving, copy creation, data synchronization, export (for backup / sharing), import (restore chat data).
- New Features : Support for browser local storage, with operations including data synchronization, import, export, deletion, and statistics.
- Smart max_tokens (Automatically adjusts the max_tokens)
- API key, OpenAI API balance inquiry, custom API endpoint, custom API requests (model, max_tokens, temperature, top_p, presence_penalty, frequency_penalty.).
- Markdown support (code block, syntax highlighting, displaying images, displaying tables, article formatting, etc.).
- system prompt, token statistics, one-click copy, automatic parsing of OpenAI error codes, function help.
- Undo, Retry, Clear Context, and supporting keyboard shortcuts.
- Supports English and Chinese-Simplified, custom avatars (supports Upload Avatar), custom style (font size, WeChat-style time format, message time, distinguishing multi-turn contexts, etc.).
- Single-round chat history, all chat history, and operation history can be exported in Markdown format, supporting export as
.md and .txt .
Updates
- August 28th Added Contextual mode.
- August 21st 1) Upgrade to support GPT-4 model. 2) Supports third-party API keys (reverse proxy service providers).
- July 29rd 1) Upgrade to support the latest model, including 16K models. 2) Optimized data management features: ability to display chat size, support for erasing logs, improved prompts, etc.
- July 23rd 1) New Feature: Auto Key Rotation , 2) New Feature: Automatic Retry , 3) Added more error code categories; Fixed default language modification bug; Addressed font blurriness on Windows, chat box bugs, and other style issues; Optimized user experience.
- July 9th Optimized mobile compatibility; Fixed style issues; Added page scaling functionality for mobile.
- July 5th v7.05 1) Added local storage and data management. 2) Added: multiple chats, automatic saving, copy creation, data synchronization, chat export and sharing, and chat import and restoration. 3) Bug fixes for Windows compatibility (scrollbar, dropdown menu styling errors). 4) Optimized UI.
- June 15th New Feature: Upload Avatar , Upgrade the Smart max_tokens algorithm with the feature to automatically resend requests in case of errors. Adjusted CSS compatibility.
- June 12th The new Smart max_tokens feature was added, which automatically adjusts the max_tokens parameter. Bug fix: Fix the issue that the prompt allows saving blank space.
- June 9th Added explanation for OpenAI error codes; Fixed some issues related to:relative path avatar; style during loading; some browsers not supporting fonts below 12px.
- June 7th v6.07 This is an almost brand new version with significant changes, upgraded features and improved user experience. It is applicable to more scenarios and is also very stable. It has been tested for two weeks without any errors.
Updates before June 7th
