Melaya · Build high-trust AI agents for any workflow, and put them on your phone. The visual agent builder, with Device Control as its flagship: Claude Code, GPT, or Gemini operating your real phone apps, one verified action at a time, with your approval on every action.

Getting started

Welcome to Melaya

Build AI agents, automate your phone, connect your tools, and work with a copilot - all in one workspace.

2 min

Melaya is a workspace for building and running AI agents. Instead of doing repetitive work by hand, you describe what you want, wire up the right tools and models, and let agents carry it out - end to end. Everything runs inside one place: you design the work, launch it, watch it happen live, and improve it over time.

A finished piece of automation in Melaya is called a pipeline (an agent, or a team of agents, with a goal, tools, and a model). You build pipelines in the Agent Builder, run them, and every execution shows up as a run you can inspect.

What you can do

Build agent pipelines
Design agents in the Agent Builder - give them a goal, a model, a persona, tools, and knowledge, then launch them. Pipelines can be a single agent or a coordinated team.
Automate your phone
Use Device Control to let agents operate approved apps on a paired Android phone - tapping, typing and navigating like a careful assistant, with you in control of what's allowed.
Connect tools & services
Give agents real capabilities from the Tools library and link outside accounts through Connectors, so agents can act on your actual data and services.
Chat with the copilot
The Assistant is your Melaya copilot. Ask it to help you set things up, explain features, or get work done conversationally - it's also where new users start onboarding.

The big picture: build, run, inspect

  1. 1
    Build
    Open the Agent Builder to design a pipeline - pick a model, add tools and connectors, and define what the agent should do. This is where you spend most of your time.
  2. 2
    Run
    Launch the pipeline. Each launch becomes a run grouped under a project, so your work stays organised as it grows.
  3. 3
    Inspect & improve
    Watch live activity on the Overview, follow the reasoning step by step in Traces, and measure quality in Evaluations - then refine and run again.
Not sure where to begin?
Head to the Assistant and tell it what you're trying to automate. If it's your first time, Melaya guides you through a short onboarding there. You can also open Guided tutorials at any time to learn features in context.

The Overview page is your home base once things are running. It shows your totals - Projects, Runs, Pipelines, LLM Invocations and Total Tokens (with cost) - plus an activity chart, your most Recent Pipelines, and a cost breakdown of your AI spend by model, provider, agent, project or pipeline.