Minitap Accuses Google's Artemis Project of Removing Open-Source Credit

Minitap documents disputed code history

Minitap has accused Google's Artemis mobile-automation project of incorporating parts of its open-source mobile-use software while failing to retain clear credit for the original developers. The startup published code comparisons and repository-history evidence on September 11, 2026, and said it had contacted Google's team seeking acknowledgment and corrected attribution.

The account comes from Minitap chief executive Nicolas Dehandschoewercker and has not been independently adjudicated. Minitap says its team found matching Android device-connection code, identical instructions for an agent called Hopper and a messaging example with the same names, comments and cleanup steps. It also says both projects once exhibited the same file-handling bug, which Artemis later fixed.

Minitap's post describes repository history as central to the dispute. According to the company, an earlier Artemis package file named Minitap contributors Pierre-Louis Favreau, Jean-Pierre Lo and Dehandschoewercker. A later version replaced those names with another author, with the file's author list being the only reported change. Minitap says GitHub records show that replacement followed a force push in August and that the old revision is now detached from the main branch.

Attribution, not reuse, is the issue

Minitap does not argue that nobody may build on mobile-use. It released the research project as open source and says it wants others to improve it, commercialize ideas or create competing products within the license. Its complaint is that downstream reuse should preserve the software's origin and applicable notices.

The company says the compared mobile-use code is under Apache License 2.0. That license permits broad reuse but sets conditions for redistribution, including retention of applicable copyright and attribution notices and identification of changes. Minitap said the Artemis README it reviewed did not credit mobile-use. Clear provenance also helps later users locate maintainers, understand design choices, report defects and return improvements to the people who created the upstream work.

The post also raises a separate leaderboard concern. Minitap says AndroidWorld displayed an earlier 91.4% result for mobile-use while showing Artemis at 99.1%, despite later Minitap submissions claiming 94.8% and then 100%. It explicitly notes that these results are self-reported and not independently verified, and says it has no evidence linking unanswered leaderboard emails to the repository changes.

Minitap has requested that Google acknowledge that Artemis is derived in part from mobile-use, identify the contributors and repair the attribution record. Until Google responds or the repository record is independently reviewed, the copying and removal allegations remain Minitap's documented claims rather than established findings.