Skip to main content

Rundra

Rundra is a portable experiment-execution framework for reproducible scientific computing. It provides stable run and task identities, explicit stochastic seeds, structured JSON output, immutable staged inputs, scheduler integration, and result retrieval through one command-line interface.

Installation

Rundra requires Python 3.12. Install the command as an isolated user tool:

uv tool install rundra
rundr --version
rundr help

Install the optional MCP interface with:

uv tool install 'rundra[mcp]'
rundr-mcp --help

Basic workflow

An experiment combines an executable command, a YAML configuration, explicit resources, and one or more integer seeds:

rundr validate experiment.yaml --json
rundr plan experiment.yaml --seeds 0:3 --json
rundr submit experiment.yaml --seeds 0:3
rundr wait --last --progress
rundr fetch --last

rundr run performs submission, waiting, and retrieval synchronously. Remote targets can combine SSH transport, a Slurm, OpenPBS, or HTCondor scheduler, rsync or shared staging, and an Apptainer container runtime. HTCondor currently requires an explicitly shared workspace and supports reliable vanilla Task clusters without dependencies or compact workers. Site connection, account, queue, workspace, and authentication policy remain explicit operator configuration.

If an asynchronous submission is interrupted, rundr resume RUN_ID recovers its durable scheduler outcome without creating a duplicate Run. rundr list returns compact paginated summaries, while rundr tasks RUN_ID pages through Task details. Automatic fetch can avoid redundant bulk transfer when client and target paths are safely visible through the same filesystem.

Automation

All important lifecycle commands support structured JSON. The optional MCP server exposes the same operation layer over stdio or authenticated Streamable HTTP, including submission recovery and paginated discovery. Network transport requires an environment-sourced bearer token and an explicit host policy; TLS should terminate at an operator-managed reverse proxy.

Rundra does not store SSH keys, passwords, registry credentials, or scheduler credentials. Run only experiment definitions and source trees that you trust, because their declared commands execute with the configured user's privileges.

The source repository contains development documentation, examples, schemas, and the complete test suite.

Metadata

Release files for rundra 0.1.6

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for rundra 0.1.6
File Size Uploaded
rundra-0.1.6.tar.gz 237.0 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for rundra 0.1.6
File Interpreter ABI Platform
rundra-0.1.6-py3-none-any.whl Python 3 none any Details

Total release size: 512.3 kB

Release files / rundra-0.1.6.tar.gz

Download URL rundra-0.1.6.tar.gz
Size 237.0 kB
Tags Source
SHA-256 checksum
How to use checksums
c20979d02635986017f9845c345da5b3b524fa5c82467191107ac90ed3cb9a4f
BLAKE2b-256 checksum
How to use checksums
ae547528a8f4f83c181faf8741f5e256031f056a09e99fc36824336e8ea5cc2c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 28, 2026.

Transparency log

Release files / rundra-0.1.6-py3-none-any.whl

Download URL rundra-0.1.6-py3-none-any.whl
Size 275.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
b8d54e2b429b9f3eac1be9d2e1203e2d0cc3ef14353f6a4a69854c1f76f7d51c
BLAKE2b-256 checksum
How to use checksums
a5a01c8a737ba0402ee82a2f9d16090f7a665824336f2e8030451fe3ab4ef3c6
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 28, 2026.

Transparency log

Release history Release notifications | RSS feed

0.1.7

2 release files

This release

0.1.6 This release

2 release files

0.1.5

2 release files

0.1.4

2 release files

0.1.3

2 release files

0.1.1

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page