Google builds OpenClaw alternative; Remy adds Gemini Assistant

Google is building an in-house OpenClaw alternative for model deployment and tool integration while Remy updated the Gemini app to add persistent assistant features.
Google is building an in-house alternative to OpenClaw while Remy updated the Gemini app to add persistent assistant features. Google has described the project in internal briefings and to partners, and Remy rolled assistant capabilities into recent app updates.
Google’s project is intended to provide another route for model deployment, tooling and integration. The work focuses on infrastructure and interfaces that let models connect to external tools and services, and on packaging and distributing models across cloud and device environments. Google has outlined plans to expand options beyond existing open standards and to offer tighter integration with its infrastructure.
Remy’s update adds persistent assistant functions to the Gemini app’s chat interface. The app can maintain context across interactions, store user preferences, recall earlier requests and carry out simple follow-up tasks within the app’s permission settings. The assistant can also trigger actions tied to third-party services when users grant consent.
Technical details emphasize interoperability and safety. Google’s alternative includes mechanisms to control model access to external tools, manage permissions and monitor activity to limit risky or unauthorized actions. Remy’s assistant relies on explicit user consent and the app’s existing account and permission controls to determine what actions the assistant can perform.
OpenClaw is a protocol and toolkit designed to standardize how language models call external tools and handle agent workflows. Google’s alternative aims to offer similar capabilities with closer integration into Google’s cloud and device services, giving partners another option for building tool-enabled model experiences.
Developers and enterprises can choose to extend app-level behavior, as Remy did with Gemini, or use platform-level frameworks such as Google is developing for model orchestration and tool integration. Both projects remain under active development and will change as developers test integrations and report issues.








