Jan

What is Jan?

Jan is an open-source alternative to ChatGPT that lets you use AI models either locally or by connecting to cloud models. The core purpose of Jan is to give you a personal “intelligence” assistant whose responses and context are tied to your own setup and preferences.

Jan supports multiple model sources (open models and online providers) and offers tools that connect to common work and productivity services. The product is presented as “built in public,” with development and releases shared via GitHub.

Key Features

• Model selection for local and online use: Choose from open models or plug in online models from providers listed on the site (e.g., OpenAI, Anthropic/Claude, Google/Gemini, and others).
• Connectors to email, files, notes, and calendar: Use integrations for Gmail, Amazon (shopping), Google Search, Notion, Figma, YouTube, Slack, Google Drive, and Jira as described on the site.
• Memory (coming soon): Jan’s memory feature is described as carrying over context and preferences so you don’t have to repeat yourself.
• Focused UI for brief answers: The site emphasizes a minimalist UI and use patterns like requesting brief, to-the-point responses.
• Open-source distribution and community: The project is promoted as free and open source, with GitHub releases and community links (Discord, Hugging Face) shown on the page.

How to Use Jan

• Download Jan: Get the Mac download from the Jan site (the page indicates a “Download for Mac”).
• Pick a model source: Select an open model or connect Jan to an online model provider listed in the models section on the site.
• Connect your tools (optional): Enable the connectors you want to use for tasks involving email, files, documentation, design tools, collaboration, or project management.
• Start a conversation: Ask questions directly, and—when the memory feature is available—use the same context to reduce repeated prompts.

Use Cases

• Ask for brief answers during design work: When you’re working in Figma or prototyping, you can ask Jan for short, to-the-point responses to design and workflow questions.
• Summarize or retrieve information from your workspace: Use connectors such as Google Drive, Notion, Slack, or Jira to help you find relevant information or discuss items from those tools.
• Research via web search and video content: With the Google Search and YouTube connectors listed on the site, you can ask questions that reference what you’re trying to learn or locate.
• Manage work communications and tasks: Use Gmail (and other listed integrations) to support Q&A or organization around your inbox and related work artifacts.
• Preference- and context-based assistance (future memory): For ongoing projects, use Jan’s described “memory” behavior to keep track of context and preferences once it’s available.

FAQ

• Is Jan open source? Yes. The page describes Jan as “Free & Open source” and highlights development resources like GitHub.

• Can Jan run AI models locally? The site states you can run open-source AI models locally or connect to cloud models.

• Which model providers does Jan support? The models section lists multiple providers and model families, including OpenAI, Claude/Anthropic, Gemini/Google, Llama/Meta, Mistral/Mistral AI, Qwen/Alibaba, DeepSeek, DeepSeek Gemma, and others shown on the page.

• Is there a memory feature? A memory capability is mentioned as “Coming Soon,” with the description that your context carries over so you don’t repeat yourself.

• What tools can Jan connect to? The page lists connectors including Gmail, Amazon, Google Search, Notion, Figma, YouTube, Slack, Google Drive, and Jira.

Alternatives

• Other ChatGPT-compatible desktop apps: Look for alternatives that also focus on local-first model use and/or connector-based workflows, if you want a similar “assistant on your machine” experience.
• Local AI chat tools using open-source models: If your primary goal is running models locally, consider local chat clients that let you manage open model backends without relying on a single hosted provider.
• Agent/workspace assistants with integrations: If you primarily want tool connectivity (email, docs, project tracking, design), compare assistants that emphasize connector ecosystems rather than model flexibility alone.
• Cloud-only chat assistants from model providers: If local execution isn’t required, cloud assistants from major providers can be simpler to set up, though they may not match Jan’s local/open-source positioning as described on the site.