Gemma is Google's family of open models, designed to run locally, even on a laptop or a phone. Its latest generation is under the Apache 2.0 license and supports tool calling natively: enough to drive Mailpro without a dedicated server.
Gemma has been published by Google DeepMind since February 2024, based on the work behind Gemini. Since its latest generation, in April 2026, Gemma has moved to the Apache 2.0 license, a classic open license.
Gemma comes in very small sizes, designed for a laptop or a phone, and in larger sizes that fit on a single server GPU. Tool calling, JSON output and system instructions are built in.
It is the easiest model on the list to run on an ordinary machine: ideal for a simple email marketing assistant, with no hardware investment.
| Developer | Google DeepMind (United States, United Kingdom) |
|---|---|
| License | Apache 2.0 since the April 2026 generation |
| Sizes | From very small (laptop, phone) to medium (one server GPU) |
| Tools | Native tool calling, structured JSON output |
| Strength | Runs on ordinary hardware |
No dedicated server needed for an assistant that adds contacts and sends confirmations.
Check a credit balance, add a subscriber, send a confirmation SMS.
Apache 2.0: free commercial use, with no special conditions.
A daily cap and approval for campaigns, even for a small model.
Install Gemma with Ollama or LM Studio; a quantized version runs on a recent computer.
Use an interface that supports remote MCP servers: LM Studio, Open WebUI or LibreChat.
Add the Mailpro MCP server using the server address shown in your Mailpro account, on the AI Assistants (MCP) page, with OAuth sign-in (details: self-hosted AI).
In Mailpro, on the AI Assistants page, set how it sends: with approval or direct, with a daily cap.
The smallest sizes handle one request at a time well; to chain analysis, writing and sending, choose a larger size.
Gemini is Google's online assistant; Gemma is its open model, which you install yourself.
The latest generation has moved from Gemma's own license to Apache 2.0: check the license if you use an older version.
On the Mailpro side, the rules are the same for every assistant: by default, each send waits for your approval, or goes out directly within a daily cap; a log keeps track of everything, and one click suspends all sends. The details.
Yes: the smallest Gemma sizes are made for a workstation, or even a phone, especially in a quantized version.
Yes, the latest generation supports tool calling natively; go through an interface that supports MCP servers.
Gemini is Google's online assistant; Gemma is the family of open models you install yourself.
All questions: AI assistants FAQ.
Create a free Mailpro account, connect your assistant from the AI Assistants page of your account, and stay in control of every send.
