Skip to content
v0.2.1 · Open Source · GNU GPL v3 GitHub

Liatir

Bioinformatics on your machine.
Under your control.

An open-source desktop app that runs bioinformatics tools, AI models and pipelines on your own computer.
The models install in one click instead of a day of conda, Docker and CUDA versions — and nothing you open ever leaves the machine.

No cloud or server
No conda, no Docker
100% free forever

Private by design

Stay on your machine. No API calls, no servers, no cloud, no telemetry — Liatir runs the same on an air-gapped workstation with no access to the web as on a connected laptop.

Models that simply install

Geneformer, scGPT and the rest are free to download and a day of work to install. Liatir delivers them as signed, verifiable packages: one click, no environment to build, and they run offline — on your GPU when you have one.

Everything speaks one language

Native tools, AI models, your own plugins, API calls and Nextflow workflows share the same inputs, outputs, Jobs, Results and provenance. That is why any step can feed any other — and why an assistant can drive the whole thing.

Everything in one place

From raw reads to annotated results without leaving your desktop — and every step hands the next one something it can actually read.

Tools, ready to run

FastQC, fastp, seqkit, samtools, bcftools, BWA-MEM, minimap2 and SnpEff, bundled and verified. Nothing to install and no containers to manage.

Local AI models

Single-cell embeddings and annotation, molecular relaxation and dynamics — on your own hardware, on your GPU where there is one. A model family is added only once it has been validated end to end, which is why the list grows slowly.

The workflows you already have

Already using Nextflow or similar? Liatir can run external workflows as a step like any other. Nothing to rewrite, nothing to abandon.

Visual pipelines

Connect tools, AI models and your own plugins by drawing the connection. No code required, unless you want it.

API connectors

Wire an external service into a workflow and drop it into a pipeline as an ordinary node — configured visually, not hand-coded.

Can be extended endlessly

Build anything on top of Liatir: .lia plugins can run with Node, Python, or WASM. They can run as pipeline nodes or by itself.

Built-in viewers

Inspect genome tracks, protein and molecular structures, and single-cell data inline, right next to the step that produced them.

Your assistant can run it

An AI assistant on your computer can start a pipeline you saved — that exact version, with only the files you allowed, and only after you approve the run.

Native speed

Every file read, process and parse runs in Rust through Tauri. Open a 40 GB BAM, stream gzipped FASTQ or scan a dense VCF without freezing the interface.

Run your first analysis today

Liatir is free and runs on your own computer, even without a network connection.