MoleditPyJob Manager

The computational companion to MoleditPy

Keep your
calculations moving.

From a single workstation to a research cluster. Submit calculations, share resources and bring results home — in one place, with or without an AI agent.

Get Job ManagerRead the guide
OPEN SOURCE / LGPL-3.0MOLEDITPY PLUGIN + STANDALONEWINDOWS · macOS · LINUX
Job Manager / Job MonitorYOUR CALCULATIONS, IN VIEW
Job Manager with fictional running, waiting and completed calculations, signed priorities and the Tail File button
Actual application UI. All jobs and hosts shown are fictional examples.01 / JOB MONITOR

Your calculation. Your destination.

SSHSLURMPBS / TorqueSGE / UGELocal + WSLJob Server

01 / From input to insight

Less tending to jobs.
More time for the science.

A consistent workflow across your own machine, remote workstations and cluster schedulers.

01 — SUBMIT

Prepare once. Submit anywhere.

Choose a host, add inputs and reuse a command template or preset. Submit from the desktop, an input generator or the local API.

INPUTS / PRESETS / TEMPLATES
02 — MONITOR

Know what is happening.

Follow job states, inspect logs and check host CPU and memory. Get notified when a calculation finishes.

JOB STATES / LOGS / NOTIFICATIONS
03 — COLLECT

Bring the results back.

Download outputs automatically and open them with an installed MoleditPy result analyzer. Keep or export your job history.

AUTO DOWNLOAD / ANALYSIS / EXPORT

02 / Built for agent workflows

Many agents.
One shared resource budget.

Let agents submit work through the local API. On a workstation, the shared helper queue schedules calculations within the CPU and memory budgets you set, helping prevent a flood of submissions from overloading the machine.

8 cores + 32 GB budgetTwo jobs requesting 4 cores and 16 GB each run together. Additional jobs wait for capacity.
Explore the API
Shell / agent workflow
# Submit, wait, fetch results
python -m job_manager.api_client submit molecule.inp \
    --host workstation \
    --command "/opt/orca/orca {input} > {stem}.out" \
    --cpus 4 --memory 16GB \
    --wait

# Exit 0 if DONE, else 2
Local HTTP API · Token found automatically · Python client for scripts

Use the same host profile and helper queue, with accurate resource requests. Budgets schedule declared requests; they do not enforce process limits or account for unrelated applications. Force run bypasses budgets. Cluster destinations use their own scheduler. The API is opt-in, loopback-only, and requires Job Manager to be running.

Actual Job Manager Host Monitor displaying fictional workstation and cluster CPU and memory statistics
Actual application UI with fictional host names and sample readings.

03 / A view of the bigger picture

Keep an eye on
the whole machine.

The Host Monitor puts CPU, memory and active jobs side by side. See where your work is running before deciding what to submit next.

  • Desktop host monitoring and history graphs
  • Optional read-only browser dashboard
  • Desktop notifications and chat webhooks

04 / Fits the way you work

A companion, wherever you compute.

01 / CONNECTIONS

Across machines and platforms

Use OpenSSH or paramiko for remote hosts, run locally with bash or PowerShell, connect to WSL, or use the LAN Job Server.

Connection guide
02 / SCHEDULING

Put the queue to work

Set CPU and memory budgets, chain dependent jobs, pause the helper queue or schedule a later start. Cluster jobs use the scheduler already in place.

The helper queue is a small script copied to the host and run there, so jobs keep starting in order after you close Job Manager, log out or lose the connection. Results download the next time Job Manager runs.

How the helper works
03 / YOUR WORKSPACE

In MoleditPy. Or on its own.

Launch as a MoleditPy plugin or a standalone desktop app. Use the same hosts, job history and settings, and return to the editor for result analysis.

Standalone setup

05 / Ready for your next calculation

Give your jobs
a place to land.

Free and open source. Start with the MoleditPy plugin, or run the desktop app independently.

Open Plugin Explorer Releases

  1. Install Job Manager

    In MoleditPy 5.x, open the Plugin Installer and install Job Manager from the official plugin collection. Or install the standalone desktop app with pip, no MoleditPy needed.

  2. Connect a host

    Open Extensions → Job Manager → Job Monitor → Hosts. Choose a backend and scheduler, configure resources, and test the connection.

  3. Submit your first calculation

    Choose New Job, add your inputs and check the script preview. For automation, enable Local API and follow the API guide.

Standalone, without MoleditPy: install from PyPI and launch.pip install moleditpy-job-manager
moleditpy-job-managerPython 3.10+. Add the [ssh] extra for paramiko or [job-server] for Job Server hosts. It shares hosts, job history and settings with the plugin. See Job Manager without MoleditPy.