RigSpark
Hardware-aware CLI that checks if local models fit your hardware and estimates speed.
Tested on Linux
Use it
Linux x86_64
Download prebuilt archive and verify checksum
mkdir -p /tmp/rigspark-dl && curl -fL -o /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu.tar.gz https://github.com/shashankswe2020-ux/rigspark/releases/download/v3.0.6/rigspark-x86_64-unknown-linux-gnu.tar.gz && echo "d139e9f656e26af006332201810ed0c8b8f21b54c5987c4bc12f24d453d4efae /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu.tar.gz" | sha256sum -cExtract archive and install binaries to PATH
tar xzf /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu.tar.gz -C /tmp/rigspark-dl && cp /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu/rigspark /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu/llmup /tmp/rigspark-dl/rigspark-x86_64-unknown-linux-gnu/rigspark-gui ~/.local/bin/ && chmod +x ~/.local/bin/rigspark ~/.local/bin/llmup ~/.local/bin/rigspark-gui && rigspark --version
Check that it works
rigspark --versionThe output should contain “rigspark 3.0.6”.
First thing to try
rigspark recommendRun rigspark recommend to see which models fit this machine.
What it does
- Get a yes, slow or no verdict on running a modelTested
How to use it
- Run
rigspark recommendto rank models for this machine with a verdict for each. - Run
rigspark can-run <model>to check one model before you download it.
- Run
- Estimate tokens per secondNot shown working
- Pull, verify and serve modelsNot tested
- Generate images and short videosNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 1 of 4 features work
- Works offline
macOS and Windows: the project says it runs there; pickfit hasn’t tried it.
Good for
- rank models for this machine
- check one model before download
- chat with the active model
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 31 Oct 2026
- Version 3.0.6
Differences the test machine hit
- README: cargo binstall rigspark-cli / cargo install rigspark-cli (needs Rust toolchain)
In the test: downloaded prebuilt rigspark-x86_64-unknown-linux-gnu.tar.gz from GitHub releases, verified SHA-256 and copied rigspark/llmup/rigspark-gui to ~/.local/bin - README: rigspark can-run llama3.1:8b
In the test: can-run exits 1 when the verdict is no; judge by output text not exit 0 - README: offline yes/slow/no verdicts for this machine
In the test: container reports 0.0 GiB usable memory with GPU-probe fallback warnings, so every verdict here is no/ram-bound; same commands give yes/slow on real hardware