Working on ML experiments with multiple collaborators often means dealing with multiple notebook runs, Kaggle accounts, and output files.
I wanted to make that process less repetitive, so I built KaggleFleet.
The idea
Instead of manually switching between Kaggle accounts and tracking notebook execution separately, KaggleFleet gives you one local interface to manage the workflow.
You can keep your notebooks in your local project, select an account, choose an accelerator, push the notebook, and monitor its execution.
For collaborative research teams, it supports multiple independently authenticated co-author accounts and simultaneous notebook execution across them.
Features
1. Local-first notebook workflow
Keep your notebooks in your local project and use your preferred editor.
2. Multiple accounts
Configure co-author accounts independently through Kaggle's browser authentication flow.
3. Parallel execution
Launch experiments using different accounts without manually switching between them.
4. Centralized monitoring
Use the local web panel to see accounts, running jobs, statuses, and logs.
5. Output collection
Download completed kernel outputs into your local project.
6. AI-agent integration
Commands support JSON output and meaningful exit codes, making it easier to integrate with coding agents and scripts.
Try it
pipx install git+https://github.com/towhaEL/KaggleFleet.git
or
pip install kaggle-fleet
kfleet doctor
kfleet account login --label "Me"
kfleet notebook list
kfleet run push train.ipynb --watch
kfleet serve --open
What's next?
I'm looking for early users, particularly ML researchers and people working on collaborative Kaggle projects.
I'd love to learn what parts of the workflow are still painful and what features would make this genuinely useful.
Top comments (0)