Presenting Minitap Platform!

You can now deploy and observe mobile AI agents with one line of code! No local setup, no GitHub repos needed anymore: mobile-use runs directly in the cloud. minitap is now a full platform where you can:

• Execute real mobile tasks end-to-end
• Create and customize your own mobile AI agents
• View complete observability: task logs, costs, agent behaviors, and how agents interact

With everything now centralized at minitap.ai/platform , it’s never been easier to build with mobile-use.

Here’s what it looks like, and why it’ll x10 your speed of development!

First, welcome!

The features:

1. Saved automation templates

The easiest way to save and track the way you want your AI agents to interact is on the Automations section. All you have to do is prompt the automation: for example, “Retrieve all employees working at minitap on LinkedIn”. The template is saved, and the task ready to run 🚀 never been faster or better!

2. Your personalised LLM configurations

We know different agent configurations achieve precisely what you want. Now, you have full control: do you want Gemini 2.5 Pro as your cortex? You got it. You prefer Llama 4 Maverick as your orchestrator? Perfect - just select it. You own the ingredients to cook that configuration up however you want it to look like.

And of course, you can save your carefully curated configurations to reuse whenever.

3. Complete obsevrability on SDK

Agents don’t run in a blackbox. You can see exactly how your tasks run on SDK: including agent behaviour, time, and chain-of-thought. By knowing all details of the execution timeline, you can retrace tasks, adjust agents, and see what works.

Minitap isn’t only a repo. It’s a platform: your platform to execute fast, and execute well, with full clarity.

Here’s our demo video :‍

‍

Technical contributions by :