Hello Google AI Developer Community,
We’re developing Zirkle, a local-first, user-controlled AI conversation manager for Windows and Android. We have working integration prototypes and are seeking clarification about officially supported Gemini integration methods.
We’ve reviewed existing discussions distinguishing the Gemini API from the consumer Gemini application. Our questions specifically concern the consumer experience, rather than building a separate chatbot using the Gemini API.
1. MCP integration: Can third-party applications register remote MCP connectors for use in the consumer Gemini web and Android apps? If so, what are the documented requirements?
2. Memory retrieval: Can Gemini autonomously invoke an authorized connector to retrieve relevant information from a user’s local encrypted archive during normal conversations, without requiring an explicit command each time?
3. Conversation preservation: Does Google provide a supported API, event mechanism, or other authorized method for preserving conversation content from Gemini’s consumer web and Android applications as conversations occur?
4. Local-first architecture: Can a secure, authenticated bridge between Gemini and an on-device database be supported through MCP or Android App Functions?
We are not requesting access to Google’s internal systems. Our objective is to identify supported integration methods and respect user consent and platform security requirements.
If these questions fall outside the Gemini API team’s scope, could a Google developer representative direct us to the appropriate consumer Gemini integrations team or official technical contact?
Website: https:/myzirkleapp.com
Thank you,
Zirkle Development Team