While developing the Meru AI MVP for launching it on the testnet for beta testing on the web, we decided to create a proof of concept (PoC). This PoC exemplifies our dedication to delivering cutting-edge technology. Its purpose is to offer investors and enthusiasts a clear insight into the capabilities of Meru AI and encourage them to explore its potential firsthand. By presenting this PoC before the $MERU TGE, we aim to engage advisors and venture capitalists more effectively. As we continue to work on the actual MVP on the testnet, this PoC serves as a small demonstration of the boundless possibilities of Meru AI. Additionally, the PoC is fully open-source on our official GitHub organization account.

This proof-of-concept script runs entirely on your local machine. It allows you to:

  1. Run the script with a single click.
  2. Automatically reply to emails in your Gmail account.
  3. Customize the tone of the auto-reply based on Gemini & PaLM 2 prompts.
  4. Set rules or filters to reply only to specific emails.
  5. Move auto-replied emails to a separate inbox for easy tracking.
  6. Ensure your data privacy - the tool runs locally on your PC and does not collect any data. Once you stop the program, it disappears along with any email/data it has read.

PoC GitHub Repository

Tech Stack

  • Python3
  • Pip3
  • Langchain
  • Gemini Model (Google AI)
  • PaLM 2 Model (Google Cloud Vertex AI)
  • IMAP protocol

Gemini VS. PaLM 2

The latest update to this PoC now supports switching between using the Google AI (Gemini) language model and the Google Cloud Vertex AI (PaLM 2) language model.

https://x.com/meru_ai_app/status/1777122314355671517

Utilizing Gemini is free of charge. Utilizing PaLM 2 entails costs and necessitates setting up billing on your Google Cloud account, as Google is deprecating the PaLM 2 API while maintaining access to this API through Vertex AI. Reference: Google’s PaLM Deprecation Notice. For PaLM 2, it is distinct from the Google PaLM integration. Google has opted to provide an enterprise version of PaLM through GCP, supporting the models made accessible there. By default, Google Cloud does not employ customer data to train its foundation models as part of Google Cloud’s AI/ML Privacy Commitment. Additional information on how Google processes data can be found in Google’s Customer Data Processing Addendum (CDPA).

Visit app.meruai.net

We have added a new subdomain where our PoC will be presented. This will encourage users to test the PoC and explore Meru AI beyond the website landing page.

Navigate to app.meruai.net now. Meru AI App

Coming Soon (Already Came)

We are planning to launch a macOS desktop application for this PoC. The application will be accessible to $MERU token holders who hold a specific amount of tokens.

https://x.com/meru_ai_app/status/1777127621798604930

Meru AI PoC is now available on macOS!

https://twitter.com/meru_ai_app/status/1777529519940706705