Jan

What is Jan?

Jan is an open-source alternative to cloud-based AI assistants like ChatGPT, designed to run entirely offline on users' computers for maximum privacy and data control. It allows users to download and utilize powerful open-source large language models (LLMs) such as Llama3, Gemma, and Mistral directly from HuggingFace, enabling tasks like chatting, summarizing, and coding without an internet connection. The platform also supports integration with cloud-based models from providers like OpenAI and Anthropic for users needing additional computational power. Jan’s mission is to empower users with a local-first AI solution, prioritizing data ownership and customization through a community-driven, open-source approach under the Apache 2.0 license. By providing an OpenAI-compatible local API server, it enables developers to integrate AI capabilities into other applications seamlessly. Jan caters to users concerned about privacy, developers building AI-powered tools, and those seeking a customizable, offline AI experience.

Jan's Core Features

• Run open-source LLMs like Llama3, Gemma, and Qwen locally on your device for private, offline AI interactions.
• Set up a local OpenAI-compatible API server at localhost:1337 to power other applications with AI capabilities.
• Download and manage AI models directly from the HuggingFace Model Hub for easy access to powerful LLMs.
• Connect to cloud-based models from providers like OpenAI, Anthropic, or Groq for enhanced performance when needed.
• Create customizable AI assistants tailored for specific tasks like writing, coding, or document summarization.
• Interact with PDFs and documents to generate summaries, answers, or insights directly on your device.
• Use extensions to add cloud AI connectors, tools, or data integrations for a personalized experience.
• Adjust alignment, moderation, and censorship levels to suit individual user preferences.
• Operate fully offline with no internet required after model downloads, ensuring data privacy.
• Support for multiple platforms, including Windows, macOS, Linux, and NVIDIA GPUs or Apple M-series devices.