Jan
by Menlo Research
At a glance
Open-source desktop app for running open language models locally as an offline ChatGPT alternative.
open-source local-llm desktop privacy
- Licence
- Apache-2.0 (main; confirmed tags: Apache-2.0 at v0.5.18-rc1-beta and v0.6.0, AGPL-3.0 at v0.5.16 and v0.5.17; other versions not individually checked)
- Category
- Platforms & Infrastructure
- Business model
- Open (open weights/source)
- Availability
- Global
- Launched
- 2023
Overview
Jan is a platforms product from Menlo Research: an open-source desktop app that runs open language models on your own machine as an offline alternative to hosted chat assistants. You use it as a desktop application on Windows, macOS or Linux, and it also exposes an OpenAI-compatible local API server on localhost:1337 that other apps and code can call. No material change was found in this review. On licensing: the repository's current LICENSE (main) places the Jan app under Apache-2.0, while GitHub's licence detector reports that file as unrecognised (we did not verify why); reading the LICENSE at individual tags, v0.5.16 and v0.5.17 are AGPL-3.0 while v0.5.18-rc1-beta and v0.6.0 are Apache-2.0, but other historical versions were not checked individually — so confirm the terms for the version you install, and note that model weights downloaded separately keep their own upstream licences.
Compare this tool Suggest a correction
Key features
- One-click download and run of open models
- Fully offline chat with local files staying private
- OpenAI-compatible local API server
- Optional cloud model connections and extensions
Use cases
- Private, offline AI chat on personal hardware
- Local development against an OpenAI-style API
- Experimenting with open models without a GPU cloud
Pricing
Jan is free and open source; users optionally pay third-party providers if they connect cloud models.
Record details
Provenance, dates and links are kept here so the main sections stay about the product itself.
Links, dates and provenance
- Record updated
- 2026-10-13
- Documentation
- www.jan.ai ↗
- Founded
- 2023
- Headquarters
- SG
- Canonical URL
- https://globalaiproductindex.com/products/jan/
Selection profile
Every field says where its value comes from. "Unknown" means we have not confirmed it — it does not mean the product lacks the feature.
Selection profile11 fields1 unknown
- Tasks it fits
- Running models yourself Questions over your own documents
Mapped from the record: local chat with local files staying private, plus the documented OpenAI-compatible local server.
Suggest a correction - How you use it
- Desktop appSuggest a correction
- Runs on
- Windows macOS LinuxSuggest a correction
- Where it runs
- Local machineSuggest a correction
- Works offline
- Yes
Local inference via llama.cpp means chat keeps working without internet once a model is downloaded; the README does not use the word "offline" itself.
Suggest a correction - Official service region
- Global
Open-source desktop app with no regional service restriction; downloads are served globally.
Suggest a correction - Public API
- YesSuggest a correction
- Licence
- Apache-2.0
Read the LICENSE text: Apache-2.0 on main (shortened notice), canonical Apache-2.0 at v0.5.18-rc1-beta, and full AGPL-3.0 at v0.5.16/v0.5.17. Observation: GitHub's licence endpoint returns NOASSERTION for this repository, and we did not verify why. NOASSERTION only means the detector could not match the file to a known licence template, so it identifies no licence — a limitation of that signal, not evidence of a custom licence.
Suggest a correction - Commercial use
- Permitted
Apache-2.0 is a permissive licence that allows commercial use; no separate commercial restriction is published.
Suggest a correction - Interface languages
- Not recorded
No vendor page read in this pass states the product's own interface languages; the key exists so the field can be answered later rather than left absent.
Suggest a correction - Billing
- Open source Free Free allowance: Yes Subscription required: No
Prices are the vendor’s advertised entry price and exclude taxes; always confirm on the vendor’s own page.
Suggest a correction
Where each value comes from
- Tasks it fits
- Our inference
- How you use it
- Verified 2026-10-09 github.com ↗
- Runs on
- Verified 2026-10-09 github.com ↗
- Where it runs
- Verified 2026-10-09 github.com ↗
- Works offline
- Our inference
- Official service region
- Our inference
- Public API
- Verified 2026-10-09 github.com ↗
- Licence
- Verified 2026-10-13 raw.githubusercontent.com ↗
- Commercial use
- Our inference
- Interface languages
- Unknown
Key facts and their evidence
Each fact lists the conditions it holds under, the page it was read from and the day it was checked.
-
Licence: Apache-2.0 Verified 2026-10-13 raw.githubusercontent.com ↗
Applies to: The Jan application code in janhq/jan: Apache-2.0 on main (shortened notice). Among the individually confirmed tags, Apache-2.0 at v0.5.18-rc1-beta and v0.6.0 (canonical Apache-2.0 text) and AGPL-3.0 at v0.5.16 and v0.5.17; other historical versions were not checked individually. Does not cover model weights, which are downloaded separately.
Read from the LICENSE file itself, not from a classifier: main carries a short Apache-2.0 notice ("Licensed under the Apache License, Version 2.0"). Observation: GitHub's licence endpoint returns NOASSERTION/"Other" for this repository, and we did not verify why. NOASSERTION only means the detector could not match the file to a known licence template, so it identifies no licence — a limitation of that signal, not evidence of a custom licence.
Suggest a correction -
Platform support: Windows (.exe), macOS (.dmg), Linux (deb, AppImage, Arm64) Verified 2026-10-09 github.com ↗
Suggest a correction -
API: OpenAI-compatible local API server on localhost:1337 Verified 2026-10-09 github.com ↗
Suggest a correction -
Deployment: Runs open models locally on the user machine via llama.cpp Verified 2026-10-09 github.com ↗
Suggest a correction -
Licence: Not covered by the repository LICENSE — downloaded model weights keep their own upstream terms Our inference github.com ↗
Applies to: Scope statement: the janhq/jan LICENSE governs the Jan application code; open models Jan fetches at runtime are separate artefacts under their own licences.
The repository LICENSE covers the app only. The terms of the model weights a user downloads through Jan were not read in this review, so they are left unasserted rather than folded into the app's Apache-2.0 licence.
Suggest a correction -
Documentation: Documentation covers a desktop app quickstart and an agent quickstart Verified 2026-10-10 www.jan.ai ↗
Suggest a correction
Frequently asked questions
Is Jan free?
Yes, Jan is a free, open-source desktop application.
Does Jan work offline?
Yes. Jan runs models locally so chats work without internet and data stays on-device.
Which platforms does Jan support?
Jan runs on Windows, macOS, and Linux.
Who should not choose Jan?
Anyone who wants managed cloud inference or a hosted model service should not choose Jan: it runs models on your own hardware, so results depend on your machine, and a model must be downloaded before you can chat.
What licence is Jan released under?
The Jan application code on the main branch is Apache-2.0. Reading the LICENSE at individual tags, v0.5.16 and v0.5.17 are AGPL-3.0 and v0.5.18-rc1-beta and v0.6.0 are Apache-2.0; other historical versions were not checked individually. Model weights you download through Jan keep their own upstream licences.
Similar products
- LM Studio — Desktop application for discovering, downloading, and running open large language models locally.
- Ollama — Open-source tool for running open large language models locally on your own machine with a simple command-line interface.
- Open WebUI — Extensible, self-hosted AI interface that runs offline and connects to local or remote models such as Ollama and OpenAI-compatible APIs.
- vLLM — Open-source inference and serving engine for large language models, known for high-throughput PagedAttention.
- Command R — Cohere's open-weight LLM family optimized for retrieval-augmented generation and tool use; succeeded by Command A as Cohere's flagship.
- Dify — Open-source platform for building LLM applications with visual agent workflows, RAG pipelines, and model management.
All Jan alternatives → · All Platforms & Infrastructure products →
Related comparisons
Related guides
Sources
- Official website ↗
- Source code (GitHub) ↗
- LICENSE — janhq/jan (Apache-2.0) ↗
- LICENSE text (raw) — janhq/jan, main branch ↗
- LICENSE — janhq/jan v0.5.17 (GNU AGPL-3.0) ↗
- LICENSE — janhq/jan v0.5.18-rc1-beta (Apache License 2.0) ↗
- Documentation ↗
This record was last reviewed on 2026-10-13.
Machine-readable record: /api/products/jan.json · Spot an error? Suggest a correction