Open Source

OpenWork, an opensource Claude Cowork alternative, is silently relicensing under a commercial license

The popular local AI agent framework changed its license without announcement, limiting open-source use.

Deep Dive

OpenWork, a project built as a locally hosted alternative to Anthropic's Claude Cohere using open-source models, has triggered controversy by quietly changing its licensing terms. Originally released under the permissive MIT license, the framework has now relicensed some components under a commercial license and modified the overall project license to limit its reach. The changes were discovered by the community in a GitHub commit, notably described with a likely AI-generated message that omitted any mention of the significant licensing shift. This lack of transparency has frustrated developers who adopted the tool under its original open-source promise.

The move highlights the ongoing tension in the open-source AI ecosystem between sustainable funding and community trust. While many developers acknowledge the need for projects like OpenWork to secure income streams, the silent nature of the relicensing—buried in a commit without announcement—has been criticized as a breach of trust. The incident, detailed in GitHub issue #1412, raises questions about the future of "open-source-washing," where projects attract users with permissive licenses before pivoting to commercial models. For developers, it underscores the importance of carefully auditing dependency licenses, especially for critical AI infrastructure tools.

Key Points
  • OpenWork, an open-source Claude Cohere alternative, silently changed from MIT to a commercial license.
  • The licensing changes were discovered in a GitHub commit with an AI-generated description that omitted them.
  • The move sparks debate on sustainable open-source funding versus transparent community governance.

Why It Matters

Developers must audit AI tool licenses as silent changes can disrupt projects and breach trust.