janhq

Jan open source ChatGPT alternative: offline local AI Jan open source ChatGPT alternative: the janhq/jan repository calls itself an « Open-source ChatGPT replacement », a desktop app that downloads and runs language models on your own machine « with full control and privacy ». It is released under the Apache 2.0 licence by Menlo Research. The README lists Windows, macOS and Linux, and the documentation speaks of full offline support for the desktop app. For an SRE or a developer, the appeal goes beyond chat. Jan bundles an OpenAI-compatible server on localhost:1337, an MCP host and a command-line tool. That makes it a local inference endpoint for your scripts and agents. What does the repository actually say, how do you install it, and what can you measure on a tool that runs on a workstation? A note on honesty: span.observer has not installed or run Jan. The commands and settings come from the README, from the official documentation on jan.ai and from the repository’s docs/ files, read on 31 October 2026. The screenshots below reproduce their text. We could not read the star count from the source that day, so we do not quote it. Jan open source ChatGPT alternative: what the repository says - Licence: the repository’s LICENSE file is the Apache License 2.0, with the line « Copyright 2025 Menlo Research ». The README confirms « Apache 2.0 ». - Version: the main branch carries version 0.8.8 in src-tauri/tauri.conf.json. The repository’s changelog dates 0.8.8 to 28 October 2026. It is a security release: chat links open in your browser, and the local server checks the origin of browser requests. - Engine: the README thanks llama.cpp, Tauri and Scalar. The documentation says the llama.cpp engine is built into the app and updates with it. On Apple Silicon Macs, MLX models are supported too. - Listed features: local models from Hugging Face (Llama, Gemma, Qwen, GPT-oss), connections to cloud models (OpenAI, Anthropic, Mistral, Groq, MiniMax), custom assistants, an OpenAI-compatible API and MCP integration. One more point: « offline » applies to local models. The documentation states that requests routed to a remote provider are not offline and do not stay on your computer. A tour of the repository and its documentation The first screenshot reproduces, word for word, the README’s « Features » section and its licence. It shows the scope of the product in six lines. !Excerpt from the janhq/jan README: the six listed features (local models, cloud integration, assistants, OpenAI-compatible API on localhost:1337, MCP, privacy) followed by the Apache 2.0 licence Source : github.com/janhq/jan/blob/main/README.md The documentation then describes the local server. The defaults are host 127.0.0.1, port 1337 and prefix /v1. An optional API key goes in the Authorization: Bearer header. The second screenshot gathers the sample request and these settings. !Jan documentation: a curl request to http://127.0.0.1:1337/v1/chat/completions with an Authorization Bearer header, then the list of local server settings (host, port, prefix, API key, trusted hosts, CORS, verbose logs) Source : jan.ai Step-by-step installation Step 1: download the app The README offers one installer per platform (jan.exe for Windows 10 and later, jan.dmg for macOS, jan.deb and jan.AppImage for Linux) from jan.ai or the GitHub Releases page. Windows is also available through the Microsoft Store. On Linux, the documentation recommends Flatpak: flatpak install flathub ai.jan.Jan flatpak run ai.jan.Jan For a .deb package, the documentation gives sudo apt-get install ./jan-linux-amd64-{version}.deb, where {version} stands for the number you downloaded. For the AppImage, run chmod +x on the file, then execute it directly. Step 2: download a model Open the Hub in the sidebar, search for a model, click Download, then pick it in a conversation’s model selector. For macOS, the README gives a memory guide: 8 GB of RAM for 3-billion-parameter models, 16 GB for 7 billion, 32 GB for 13 billion. Step 3: start the OpenAI-compatible server 1. Go to Settings > Local API Server. 2. Choose the Default Model, a llama.cpp or MLX model you have already downloaded. 3. Click Start Server, then wait until the status shows the server is running. 4. Test it with the curl request in the screenshot above, replacing YOUR_MODEL_ID. According to the documentation, you read the list of available models with GET /v1/models. If no key is configured, omit the Authorization header. Step 4 (optional): the command line The documentation says the jan CLI is installed automatically the first time you launch the app. It serves a model without the interface, with jan serve, or starts an agent wired to a local model, with jan launch. Note that the CLI exposes the model on localhost:6767/v1, not on the app server’s port 1337. The third screenshot shows the CLI help and the layout of the data folder. !Jan documentation: help output of the jan command (serve, launch, threads, models, examples jan serve and jan models list), then the data folder layout with logs/app.log, llamacpp/models and threads Source : jan.ai Limits and open questions - Hardware: the documentation says the bundled llama.cpp engine does not load models on Intel Macs since 0.8.5; MLX also requires Apple Silicon. - Network exposure: choosing host 0.0.0.0 makes the server reachable from the local network. The documentation then recommends an API key and advises against leaving it empty. Trusted Hosts add a further layer of control. - Tool execution: the « Execute Tools on Server » option, disabled by default, runs MCP tool calls received by the API on the server side. Leave it off unless the client needs it. - Models and MCP: the documentation warns that not every model handles tool calls well. MCP-specific risks are covered in our guide to securing an MCP server. - Telemetry: the privacy page describes usage statistics subject to your choice at first launch, changeable later in settings, sent to PostHog (EU hosting). It states that conversation content is not tracked. Read it again before an enterprise rollout. - Pace of change: three releases in three weeks (0.8.6, 0.8.7, 0.8.8, going by the changelog dates). Test each update before rolling it out widely. Monitoring Jan on a workstation or test server Jan is a desktop app, not a service designed for Kubernetes. The project documents neither Prometheus metrics nor OpenTelemetry export for the app, so we assume none. What the documentation does provide: - Logs: logs/app.log in the data folder, with the [APP], [SERVER] and [SPECS] categories. « Verbose Server Logs » are enabled by default and detail the requests the local server receives. - Availability: a synthetic check on GET /v1/models confirms the server answers and the expected model is present. A 401 means a missing or wrong key, a 404 an unknown model ID, and « Connection Refused » a stopped server or wrong port. The documentation lists these three cases. - Resources: memory and CPU or GPU use of the process can be measured with host tools. That is where you see a model too large for the machine. The « Request Timeout » setting bounds the wait for a model response. For client-side calls, instrument the script that queries the endpoint: one span per request, with the model name and the duration. The approach is detailed in monitoring a local LLM with Ollama, and it applies to any local OpenAI-compatible server. To choose the model itself, see the operational criteria for an LLM. And Dynatrace? Jan documents no Dynatrace integration, and we do not invent one. OpenTelemetry instrumentation of the client that calls the local server remains possible with any OTLP-compatible backend, Dynatrace SaaS included; check the setup in the Dynatrace documentation. Best practices - Keep the host at 127.0.0.1 while no other machine needs the API, and set an API key as soon as you go beyond that. - Leave « Execute Tools on Server » off unless you have a specific need, and limit connected MCP servers. See also guardrails for AI agents in production. - Keep Jan up to date: version 0.8.8 fixes issues reported through GitHub Security Advisories. - Watch the app.log file after every model or version change. - If you need an engine for production, compare it with vLLM and its metrics. Frequently asked questions Does Jan really work offline? Yes for local models, once downloaded: the quickstart describes local models as « private, offline, no API key ». Cloud models and requests routed to a remote provider need a connection. Can Jan replace the OpenAI API for my scripts? The documentation presents it as a drop-in replacement for private applications. Point your client at http://127.0.0.1:1337/v1. span.observer has not tested the exact compatibility of each parameter, so check the ones your code depends on. Which licence, and can I use it at work? The code is under Apache 2.0, a permissive licence. Each downloaded model has its own licence: read it before any professional use. Conclusion Jan is a simple way into local AI. Its OpenAI-compatible server and CLI make it useful for trials, agent prototypes and environments without a cloud. Monitoring remains yours to build around it: logs, synthetic checks and host measurements. The project moves fast, so reread the documentation before each rollout. span.observer is an independent community blog, not affiliated with Jan, Menlo Research or Dynatrace LLC. Trademarks belong to their owners. Tools change quickly: check commands and options in their official documentation. --- Originally published on span.observer. French version: span.observer.

GitHub: janhq/jan download for Windows, macOS or Linux: jan.ai

聊天不想联网，也不想被云端记一笔？ Jan 是开源 ChatGPT 平替，本机跑 Llama、Qwen、GPT-oss，也能接 OpenAI、Claude 等云模型，带本地 API 服务器。 适合想离线聊的人，本地模型吃配置，先试小模型。 你先选本地还是云端？ #本地AI #聊天 @grok janhq/jan

Imagine being an oligarch thinking you can paywall AI janhq/jan

Jan: Open-source ChatGPT replacement that runs 100% offline on your computer (45k stars) What it does All text below is quoted directly from the project README. >Jan is an open source alternative to ChatGPT that runs 100% offline on your computer. Features: - Local AI Models: Download and run LLMs (Llama, Gemma, Qwen, GPT-oss etc.) from HuggingFace - Cloud Integration: Connect to GPT models via OpenAI, Claude models via Anthropic, Mistral, Groq, MiniMax, and others - Custom Assistants: Create specialized AI assistants for your tasks - OpenAI-Compatible API: Local server at localhost:1337 for other applications - Model Context Protocol: MCP integration for agentic capabilities - Privacy First: Everything runs locally when you want it to >Jan is bringing the best of open-source AI in an easy-to-use product. Download and run LLMs with full control and privacy. Repository janhq/jan Repository stats |Metric|Value| |:-|:-| |Stars|45,163| |Forks|3.1k| |Last release|v0.8.8 (2026-10-28)| |Languages|Rust 56.5%, TypeScript 39.1%, Python 1.5%| |Contributors|179| |AI contributor accounts (Claude / Codex / Cursor / Copilot)|@Copilot (19 commits), @claude (18 commits)| |AI co-authored commits|Claude: 19, Copilot: 21| Main contributor |Field|Value| |:-|:-| |Name|Faisal Amir| |GitHub|@urmauur| |Commits to this repo|1,228 (top contributor)| |Bio|UI/UX Engineer| |Company / Location|Jakarta, Indonesia| |Followers|139| ────────────────────────────── 📌 If you want to browse more repos like this, we've organized them by category and rankings here: bestgithubrepositories.com 🗂️ We're also archiving repos where possible, so useful projects don't disappear when they're removed from GitHub.

1) Jan replaces ChatGPT for everyday work You download a model once and chat with it offline. On a Mac, 8GB of RAM runs small 3B models, good for drafts and summaries. 16GB gets you 7B models. janhq/jan

janhq / jan: Jan is an open source alternative to ChatGPT that runs 100% offline on your computer. ★45164 janhq/jan

LM Studio vs Ollama vs Msty vs Jan in 2026 (Best Local AI App for Your Computer) --- title: "LM Studio vs Ollama vs Msty vs Jan in 2026 (Best Local AI App for Your Computer)" published: true description: "Off Grid AI runs local models free, with images, voice and a phone app; Pro is $79.99 once. LM Studio, Ollama, Msty and Jan run on your computer too." tags: ai, privacy, productivity, opensource canonical_url: getoffgridai.co cover_image: getoffgridai.co --- All five apps run AI on your own computer, but Off Grid AI does the most in one free download, and it is the only one with on-device models on your phone. LM Studio, Ollama, Msty and Jan are good, honest local tools. Off Grid AI adds images, voice and one API for every model for free, then memory and approved actions for $79.99 once. We make Off Grid AI. Rival facts come from their own pages, checked 30 October 2026. Download Off Grid AI · Get Pro for $79.99 once <figure> <img src="getoffgridai.co" alt="Off Grid AI Models: text, vision, image, speech and voice models to download and run on the computer." loading="lazy" /> <figcaption>The Models screen in Off Grid AI Desktop: chat, vision, image and speech models, all run locally.</figcaption> </figure> Five local AI apps side by side | | Off Grid AI | LM Studio | Ollama | Msty | Jan | |---|---|---|---|---|---| | Where models run | Your computer (llama.cpp) | Your computer (llama.cpp, MLX); hosted models on paid plans | Your computer; cloud plans for hosted models | Your computer (Ollama, MLX, llama.cpp), or hosted providers | Your computer | | Works offline | Yes, once a model is downloaded | Yes, with local models | Yes, with local models | Yes, with local models | Yes, once a model is downloaded | | Free for local use | Yes | Yes | Yes | Yes, Studio Free | Yes | | Licence | AGPL-3.0 | Not listed | MIT | Not listed | Apache 2.0 | | Account needed | No | Not listed | Not listed | Not listed | Not listed | | Computers | Mac (Apple silicon), Windows, Linux | Apple silicon Mac, x64 and ARM64 Windows, x64 Linux | macOS, Windows, Linux | Apple silicon and Intel Mac, x64 and ARM64 Windows, Linux | Windows, macOS, Linux | | Phone app with on-device models | Yes, iPhone and Android | Not listed | Not listed | Go mobile checks progress and approvals | No, desktop only | | Local API | 127.0.0.1:7878, OpenAI-compatible: chat, vision, images, speech, embeddings | "OpenAI-like endpoints" | localhost:11434, OpenAI- and Anthropic-compatible clients | Not listed | localhost:1337 | | Image generation | Yes, on Mac, Windows and Linux | Not listed | Experimental, macOS only | Media Studio | Not listed | | Voice in and out | Whisper in, Kokoro out | Local dictation in Bionic | Text to speech on Apple silicon | Not listed | Not listed | | Answers from your files | Projects, with citations | Bionic works on PDFs, decks and spreadsheets | Drag and drop text or PDF files | Knowledge Stacks | Not listed | | Shared team server | Team mode with per-person keys, free | llmster, a headless build | Cloud Team plan, $500 a month | Enterprise and Teams, by quote | Not listed | | Memory of screens, meetings and mail | Yes, with Pro | Bionic searches its own past sessions | Not listed | Not listed | Not listed | | Actions after your yes | Gmail, Linear, Jira, Notion, with Pro | Not listed | Not listed | Go agents with reviewable steps | Not listed | | Paid plan | Pro $79.99 once, or $7.99 a month, 5 devices | Bionic+ $20 a month; Pro $100 a month | Cloud Pro $20 a month or $200 a year; Max $100 | Aurum $149 per user a year, or $349 once | Not listed | | Paid plan over 3 years | $79.99 | $720 (Bionic+) | $600 (Pro, yearly) | $447 yearly, or $349 once | Not listed | Sources: LM Studio's pricing, app docs and Bionic page; Ollama's pricing, download page, docs and posts on image generation and speech models; Msty's pricing, Studio, Go and local models docs; Jan's GitHub README. All checked 30 October 2026. "Not listed" means the rival's own pages don't say. Off Grid AI's price is a limited-time lifetime offer, down from $119. What the paid plans cost over three years All five are free for local models. This table is for when you want more. | | 1 year | 3 years | |---|---|---| | LM Studio Pro | $1,200 | $3,600 | | LM Studio Bionic+ | $240 | $720 | | Ollama cloud Pro, yearly | $200 | $600 | | Msty Aurum, per user | $149 | $447 | | Msty Aurum Lifetime, per user | $349 | $349 | | Off Grid AI Pro, lifetime, 5 devices | $79.99 | $79.99 | LM Studio and Ollama mostly sell hosted compute. Msty sells features. Off Grid AI Pro sells features that run on your own hardware, so it has no credits or token limits. Pick LM Studio if… You want MLX on a Mac, a headless server build for CI, or Windows on ARM. Pick Bionic if you want a code and office-file agent with hosted open models, paid monthly. Pick Ollama if… You want a small runtime and a CLI, and you'll build your own app on top. Many tools list it as a backend. You can also keep it: Off Grid AI can use an Ollama server as a remote model. Pick Msty if… You want many local engines and cloud providers in one window, side-by-side model chats, an Intel Mac, or a managed suite with SSO for your company. Pick Jan if… You want a simple, open-source desktop chat with a local API and MCP, and nothing more. Pick Off Grid AI if… You want chat, images, voice and one API for every model in one free app, plus a phone app that runs its own models. Add Pro when you want it to remember your work. Why Off Grid AI wins here This page is about what a local AI app gives you. All five are local, so the win is the honest difference. - More in the free app. Images on Mac, Windows and Linux, voice in and out, and cited answers from your files. Ollama's images are macOS only and experimental. Jan lists none of these. - One address for every model. Chat, vision, images, speech to text, text to speech and embeddings, all from 127.0.0.1:7878. Team mode with per-person keys is free. - Your phone too. Off Grid AI Mobile runs on-device models on iPhone and Android, offline. At home it can use your computer's bigger model. No rival lists this. - Memory and actions on top. Pro remembers your screens, meetings and mail, and drafts work in Gmail, Linear, Jira and Notion that waits for your yes. - One price. $79.99 once for 5 devices, against $349 to $3,600 for the rivals' paid tiers. Honest limits. Off Grid AI's Mac build is Apple silicon only. LM Studio and Msty cover more desktop hardware, and LM Studio runs MLX. FAQ Which local AI app does the most for free? Off Grid AI Desktop. The free app runs chat, vision, image generation, speech to text and text to speech on Mac, Windows and Linux, answers from your files with citations, and serves all of it from one OpenAI-compatible API. It is AGPL-3.0 and needs no account. Which local AI app has a phone app with on-device models? Off Grid AI. Off Grid AI Mobile runs its own models on iPhone and Android, offline. Jan is desktop only. LM Studio, Ollama and Msty's pages do not list on-device phone models. Are LM Studio, Ollama, Msty and Jan local too? Yes. All four run models on your own computer, and all have a free way to do it. They differ in what comes around the model, and in their paid plans. Which local AI app works on Windows on ARM or an Intel Mac? LM Studio runs on ARM64 Windows. Msty lists Intel Macs and ARM64 Windows. Off Grid AI's Mac build is for Apple silicon only. How much do the paid plans cost? On 30 October 2026, Off Grid AI Pro is $79.99 once, a limited-time price, or $7.99 a month, for up to 5 devices. LM Studio Bionic+ is $20 a month. Ollama's cloud Pro is $20 a month or $200 a year. Msty Aurum is $149 per user a year, or $349 once. Download Off Grid AI Desktop and run every kind of model from one free app.

100% offline chat still starts with a download. Naturally. Jan lets builders run Llama, Gemma or Qwen locally on Windows, macOS and Linux instead of using ChatGPT. You need to download the models first; offline is how you run them, not how you get them. janhq/jan t.co

ChatGPT Plus Costs $20/Month. This Open-Source App Gives You a Private ChatGPT Free - Jan ChatGPT Plus: $20/month. Jan: free and open source (Apache 2.0). 🤖 45K+ GitHub stars. 1️⃣ Local models: download Llama, Gemma or Qwen from Hugging Face 2️⃣ Custom assistants for any task 3️⃣ Local OpenAI-compatible API (localhost:1337) for your other apps 🔒 Runs on your own computer, 100% offline. No account needed. With local models, your chats never leave your machine. 🔗 janhq/jan 🌐 jan.ai Note: local models need a decent computer (Jan's README, macOS: 8GB RAM for 3B models, 16GB for 7B). Follow for more free AI tools. #shorts #aitools #jan #opensource #chatgpt #localai