AI Potluck
Contribute
Back to Gap Map Organization

ClearML

company

Scores

1 product on the map — 1 open-ish.

ClearML

Openness

2 high confidence
2.0
license
Apache-2.0(the clearml SDK repository)+SSPL-1.0(clearml/clearml-server, the server the platform runs on, under a ClearML Inc copyright)
source
public(the SDK, the agent and the server are all published)

The source is fully public and the licenses are what hold the score down. The SDK is Apache-2.0, but the server the platform actually runs on carries the Server Side Public License, which the shared ladder places in competition_restricted - readable source under terms that forbid a class of use. Most-restrictive-wins applies across the license compound, and it is admissible here rather than an abstention because both names map to a declared tier. Full source plus a competition-restricted license scores 2, beside partial source, because the practical outcome matches: you can read it and not run it freely. The record carries no core-gated component: the ladder settles this product on competition_restricted plus public source before that dimension is read, so recording it would assert something the score does not rest on and no cited source establishes.

Adoption

3 high confidence
3.0

The PyPI download API reports 625,991 downloads of clearml in the trailing 30 days, inside the 100K-1M band for level 3. The declared artifact is the SDK, which is the channel users install.

Capability

3 medium confidence
3.0

ClearML decides which queue a task joins and, with dynamic GPU allocation, how many GPUs on an agent that queue commands. The agent then pulls the work, so the choice of which machine runs a given task is made by whichever agent polls rather than by a central placement decision - the same reasoning that kept hatchet out of deployment, applied consistently. Band 3 rather than higher for that reason, and the enterprise-only GPU features are counted rather than discounted.