A SvelteKit chat interface for LLMs that connects to any OpenAI-compatible API, and the codebase behind HuggingChat.
ShipBit/slickgpt
A light-weight Svelte web client for OpenAI-compatible chat APIs, using your own API key, with local chat history and shareable chats.
Open-source alternative to:ChatGPT
- Stars
- 469
- Forks
- 90
- Open issues
- 25
- Last commit
- 15 days ago
What is it?
SlickGPT is a web client for OpenAI-compatible chat APIs, written in Svelte and SvelteKit. Users supply their own API key, and chat history is stored in the browser's IndexedDB rather than on a server. The interface includes chat title suggestions, message editing and deletion, a context (system message) editor, and adjustable model parameters such as temperature and top_p.
A built-in token cost calculator shows the cost of a chat as you type and receive replies. The share feature creates a link to a chat without requiring login, and recipients can import it into their own local instance. Shared chats are stored in a Firebase Realtime Database through a serverless endpoint.
The project can run locally, be self-hosted, or be deployed to Vercel. The README lists GPT-3.5 Turbo, GPT-4 Turbo and GPT-4o, plus Mistral Large and Llama 3 via Groq, and notes that local models served by Ollama or LM Studio can be connected when self-hosting. The maintainers have declared the project end-of-life and no longer maintained.
Key features
- 01Chat history stored locally in the browser's IndexedDB
- 02Share chats via links without login, with the ability to unshare or update
- 03Editable context (system message) and adjustable model parameters
- 04Real-time token cost calculation and cost breakdown per chat
- 05Light and dark themes with a responsive, mobile-friendly layout
- 06Self-hostable and deployable to Vercel
Use cases
- Using OpenAI-compatible models with your own API key and without a subscription
- Keeping chat history private in the browser instead of an account-based service
- Self-hosting a chat interface that connects to a local model via Ollama or LM Studio
- Tracking token spending across conversations
Quick start
- 1
Clone the repository and change into its folder
- 2
Install dependencies with npm
npm i - 3
Start the development server and open it in the browser
npm run dev -- --open
Things to know
- The project has reached end of life and is no longer maintained.
- The hosted public version at slickgpt.app is offline.
- Sharing chats requires a Firebase setup, and disabling it needs code changes.
In collections
Topics
FAQ
- What is slickgpt?
- A light-weight Svelte web client for OpenAI-compatible chat APIs, using your own API key, with local chat history and shareable chats. SlickGPT is a web client for OpenAI-compatible chat APIs, written in Svelte and SvelteKit. Users supply their own API key, and chat history is stored in the browser's IndexedDB rather than on a server. The interface includes chat title suggestions, message editing and deletion, a context (system message) editor, and adjustable model parameters such as temperature and top_p.
- Is slickgpt free to use? What license does it have?
- slickgpt is open source under the MIT license. Check the license terms before using it in commercial products.
- Is slickgpt still maintained?
- Actively maintained. The last commit to slickgpt was on Oct 23, 2026.
- What can I use slickgpt for?
- Using OpenAI-compatible models with your own API key and without a subscription Keeping chat history private in the browser instead of an account-based service Self-hosting a chat interface that connects to a local model via Ollama or LM Studio Tracking token spending across conversations
- How do I install slickgpt?
- Clone the repository and change into its folder Install dependencies with npm: npm i Start the development server and open it in the browser: npm run dev -- --open
- Where can I get slickgpt?
- The source code and installation instructions are on GitHub: https://github.com/ShipBit/slickgpt
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A self-hostable Ruby on Rails ChatGPT-style chat app that works with multiple AI providers using your own API keys.
An open-source, local-first AI agent desktop client supporting MCP, Skills, ACP agents and remote control from messaging apps.
An unofficial, zero-dependency Go client for the OpenAI API, covering Responses, Chat Completions, images, audio and more.
A lightweight web GUI for ChatGPT and many other LLMs, with file-based Q&A, web search, local model deployment and an agent assistant.
A Go CLI that generates git commit messages and brief code reviews with LLMs, and can install a prepare-commit-msg git hook.