Q: Google AI Agent APK, MCP and A2A
Hi! Love your product and have been spreading the word consistently. Manu of my clients have embraced your product.
I saw an answer you gave 2 weeks ago to a question about AIAgents and MCP.
Your response wasn't entirely clear to me. However, I also have some further questions:
The recently released Google ADK is quickly moving to become the standard for AIAgent orchestration. This is further supported by the Google developed A2A protocol and finally by MCPs.
Will Your product embrace fully multi agent orchestration using Google ADK or another Agent orchestration framework?
Will the RTILA Agents communicate via A2A?
What about building MCP servers, or clients?
How will Smoking benefit from that, and what will the limitations for smoking be?
Do you have concrete plans or a roadmap concerning these points?
RTILA
May 2, 2025A: Hello Solo, thanks for your good words and suggestions
Please see a first batch of responses below:
Will Your product embrace fully multi agent orchestration using Google ADK or another Agent orchestration framework?
>>> Yes, we started working with Flowise Ai but we will look into Google ADK too. Not immediatly in the short term but mid-term we will be integrating ai Ai Chat agent inside RTILA to help with project building, command configurations etc... And also orchestrate and interconnect with external tools and agents. But our approach is to do everything self-hosted for data protection and for Enterprise clients.
Will the RTILA Agents communicate via A2A?
>>> Yes, and we are focusing/prioritizing Rest API based integrations
What about building MCP servers, or clients?
>>>
MCP will be part of whatever Orchestrator/Integration we will bring in as we are or want to be agent/tool agnostic to offer the best flexibility.
How will Smoking benefit from that, and what will the limitations for smoking be?
>>> You mean RTILA ? The main benefit is reducing the learning curve and time to setup, debug and run automations projects using natural language and avoiding as much as possible the Human discovery/building of complex CSS Selectors for web automation.
Do you have concrete plans or a roadmap concerning these points?
>>> We have a Data Centric RoadMap see this video and it will give you our vision on how we see the best use of Ai/Agents/MCPs etc..: https://www.youtube.com/watch?v=Z2S_K4CdYSk