Get LightRAG on macOS and Linux Now!
Installed with uv. Open source, with a documented one-command setup.
Get LightRAG with one command
Pick your system, reveal the step-by-step guide, copy the command, paste it into your terminal and press Enter. This command was taken from the project's own documentation and checked against it.
Not available for Windows yet.
LightRAG doesn't publish an install command for Windows anywhere we can verify, so there's nothing for us to show you. We won't invent one. It does document one for macOS and Linux — switch to that tab.
Pastes into PowerShell and installs LightRAG on your Windows machine in one step. Runs as your normal user account, and needs no administrator rights.
Not available for macOS yet.
LightRAG doesn't publish an install command for macOS or Linux anywhere we can verify, so there's nothing for us to show you. We won't invent one.
Installs LightRAG in its own isolated environment, so it can't break the Python packages you already have. Runs as your normal user account, and needs no administrator rights.
How to set up LightRAG, step by step
Open Terminal. On a Mac it's in Applications ▸ Utilities, or press Cmd+Space and type Terminal. On Linux, press Ctrl+Alt+T. Paste the command and press Enter.
Run LightRAG --help to see the available commands, and
LightRAG --version to confirm the install worked. If your shell complains about
permissions, that project usually documents a corrected command on its own site.
This installs LightRAG into its own isolated environment rather than into your system Python, which is why it won't break your other packages.
What LightRAG does
- Launching the LightRAG Server with Docker Compose
- VECTORSTORAGE: Used to store vector information for text chunks, entities, and relationships
- SUMMARYLANGUAGE: Controls the language used by the LLM when outputting entity-relation names and summaries, e.g
- MAXPARALLELPARSEMINERU: Controls the number of parallel files processed for MinerU parsing
- MAXPARALLELPARSEDOCLING: Controls the number of parallel files processed for Docling parsing
- EMBEDDINGFUNCMAXASYNC: Controls the maximum concurrency for embedding models
Summarised from LightRAG's own README.
About LightRAG
Installed with uv. Open source, with a documented one-command setup.
| Category | Command-line tool |
|---|---|
| Install method | uv tool |
| Windows | no install command published |
| macOS / Linux | documented install command available |
| Price | Free and open source |
| Command verified | Yes — checked against the project README |
| Popularity | 40k stars on the project's public repository |
On mobile
LightRAG is a command-line tool, so there is no phone app for it — a phone has no shell to run it in. You'll need a desktop, or a remote shell into a machine that has one.
LightRAG is developed by its own authors. This page is an independent reference; we are not affiliated with or sponsored by the project. The command shown here was reproduced from the project's public documentation — always check the project's own documentation before running anything.