We built this MIT-licensed npm/pip sidecar after discovering a specific architectural gap in self-hosted AI assistants like Odysseus. While these platforms do an excellent job of isolating chat history and API keys per user, they completely lack usage visibility and hard limits. We wanted to bring precise usage controls to these platforms without forcing developers to replace their existing infrastructure with heavy LLM gateways. We initially focused just on basic message metering. However, we quickly realized that to be truly useful for developers, the tool needed to integrate seamlessly into automated workflows. We evolved the project into a lean sidecar that includes both a CLI and a library with agent-native --json output. This shift allowed us to provide fail-closed monthly quota caps that work directly alongside the infrastructure you already run. Repo: https://github.com/RudrenduPaul/... MCP Servers: https://mcpservers.org/servers/r... https://glama.ai/mcp/servers/Rud... NPM: https://www.npmjs.com/package/wo... PyPI: https://pypi.org/project/workspa... We would love to hear your feedback and are happy to answer any questions!

WorkspaceGuard

Usage caps for shared self-hosted AI assistants

Usage caps for shared self-hosted AI assistants