Skip to content
Linux x86_64 build

Axam for Linux

A single tarball with everything: portable Python, the app, and all three AI models. Extract, run one script, and Axam launches in your browser — entirely offline.

Download size
~5.0 GB
Architecture
x86_64
Min RAM
4 GB
Price
Free

Tell us who you are below — then your download begins.

Axam logo

Get the Linux build

A few details, then your download begins

No spam. We use this to know who's running Axam and to send install help if you need it.

One file, everything inside

What's in the download

No second downloads, no package managers, no internet after extraction. Just one tarball.

Portable Python 3.11

No system Python needed. Won't conflict with anything you have installed.

All 3 AI models

A1 Pro (Gemma 3 4B), A2 Standard (Gemma 2 2B), A3 Lite (Qwen 2.5 1.5B). Axam picks the best one your hardware can run.

Install & launch scripts

A short INSTALL.sh sets things up. Then start_axam.sh launches the app in your browser.

System Requirements

Will Axam run on my Linux box?

Probably yes. The technical floor is glibc 2.28+ on x86_64, which covers basically every distro from 2019 onward.

Will run on

  • Ubuntu 20.04, 22.04, 24.04 LTS (and newer)
  • Debian 11 Bullseye, 12 Bookworm
  • Linux Mint 20+
  • Pop!_OS 20.04+
  • Elementary 6+
  • Fedora 29+
  • RHEL / CentOS / Rocky / AlmaLinux 8+
  • openSUSE Leap 15.2+ / Tumbleweed
  • Arch / Manjaro / EndeavourOS (recent)

Won't run on

  • Ubuntu 18.04 LTS or older (glibc too old)
  • Debian 9 Stretch or older
  • CentOS 7 / RHEL 7
  • Raspberry Pi / ARM laptops
  • Apple Silicon Macs (this is the Linux build)
  • 32-bit x86 systems

Not sure what you have? Run this in a terminal:

cat /etc/os-release | head -3

Hardware: pick your tier

Axam ships with three models and picks the best one your hardware can run. You can switch models any time from the in-app menu.

A3 Minimum
AXAM Lite

Qwen 2.5 1.5B, on CPU

  • RAM4 GB
  • Disk8 GB free
  • CPU64-bit, ~2012+
  • GPUNot required
  • Reply time3–5 sec
Recommended
A2 Sweet spot
AXAM Standard

Gemma 2 2B, on CPU

  • RAM8 GB
  • Disk10 GB free
  • CPUi5 / Ryzen or similar
  • GPUNot required
  • Reply time2–4 sec
A1 Fastest
AXAM Pro

Gemma 3 4B, on GPU

  • RAM8 GB+
  • Disk10 GB free
  • GPUNVIDIA, 4 GB+ VRAM
  • Reply time1–3 sec

Without a GPU, Axam falls back to A2 Standard.

Display
1024 × 768 minimum
Browser
Firefox 78+, Chrome 90+, Edge, Brave, Opera
Network
None — runs entirely offline

Installation

From download to chat in 4 steps

No package managers, no sudo, no system changes. Everything stays in one folder.

1

Extract the archive

Open a terminal in the folder where the tarball downloaded (usually ~/Downloads) and run:

tar -xzf AXAM-Linux-x64.tar.gz

This creates an AXAM/ folder. Extraction takes 1–3 minutes.

2

Run the installer

Move into the folder and run the install script:

cd AXAM ./INSTALL.sh

This sets file permissions and verifies the bundled Python. No sudo required — Axam stays inside the folder.

3

Launch Axam

Start the app:

./start_axam.sh

Your default browser opens automatically. First launch takes 10–30 seconds while Axam loads the AI model into RAM.

4

Start chatting

Axam picks the best model your hardware can run on first launch. Ask any question — math, science, English, history. No internet needed, no sign-in, no usage limits.

Want to switch models? Use the menu in the chat interface to pick A1 Pro / A2 Standard / A3 Lite at any time.

Optional

Verify your download

It's a 5 GB file — if your network hiccupped you'd want to know. Compare the SHA-256 of your download against the published checksum:

Expected SHA-256 AXAM-Linux-x64.tar.gz
49b9f4220c4897fe39a36ebc9628abec5ec1a188ee705f3f16d874e58dc4c7e3

Run this in the folder where you downloaded the tarball:

sha256sum AXAM-Linux-x64.tar.gz

If the output matches the expected hash above, your download is intact. If it doesn't, re-download.

FAQ

Common questions

Why is the download so big (5 GB)?

The tarball includes everything: portable Python 3.11, the Axam app, and all three local AI models. There are no second downloads, no internet calls, no package installs after extraction. The trade-off for "no internet ever" is one bigger upfront file.

Do I need to install Python or anything else first?

No. Axam ships its own Python and dependencies inside the tarball. Nothing on your system gets touched.

Does it work without internet?

Yes — that's the whole point. After you've downloaded the tarball, you never need internet again. No API calls, no accounts, no telemetry, no rate limits.

Will it run on my Raspberry Pi?

Not yet — this build is x86_64 only. ARM and Apple Silicon builds are on the roadmap.

How do I uninstall it?

Delete the AXAM/ folder. That's it. Axam doesn't write to system directories.

Is this really free?

Yes. Axam is free for educators and students. Don't pay anyone for it. If you'd like to support development, you can donate.

Need help? Stuck on install?

Email axaminfo@gmail.com with your distro version and the error you're seeing. We'll get back to you.