Product

The agent platform,
end to end.

LISA is not a collection of disconnected tools. It is one place to ask, delegate, and review — with the state of the work kept in the same account.

LISA AI agents screen
LISA AI chat screen with a real answer
LISA AI completed works screen

Actual LISA app screens

Three layers

One assistant, with room to scale.

Add complexity only when the work needs it. The conversation stays the centre of the product.

Conversation

The home screen is a chat. LISA answers with a real model response and keeps the thread in your account.

Agents

Give repeatable work a name, a role, instructions, skills, and its own model. Mention the agent to hand over the task.

Groups

Run several of your agents together on one task. LISA returns a single combined result and records the run.

Core experience

From a message to a finished run.

AI Chat

A conversation that carries the work forward.

Ask, refine, and build on what came before. LISA keeps your conversations together so you can return to the thread instead of starting over.

  • Conversation history you can reopen
  • Suggested starting points
  • Copy and reuse any answer
LISA AI chat screen showing a completed conversation with an answer from LISA

Personal AI Agents

Agents with a name, a role, and instructions.

Create agents for the work you repeat. Give each one a role, instructions, skills, and its own model. Mention an agent in a message to hand the task over.

  • Custom roles and instructions
  • Per-agent model choice
  • Mention an agent from the composer
LISA AI conversation where the prisca agent answers an API integration question

Agent Groups

Several agents, one task.

Group agents that work best together. A group run sends the task through your agents in sequence and returns a single combined result.

  • Build a group from your own agents
  • Run the group as one task
  • Review the combined output
LISA AI agent groups screen

Completed Works

A record of what was actually done.

Every agent run is recorded with its task, result, and status. Completed works is a clear archive of real output — not a feed of promises.

  • Task history with results
  • Clear completed and failed states
  • Open the work behind a notification
LISA AI completed works screen showing finished agent tasks

Model choice

Auto by default. Direct control when you want it.

Auto lets LISA choose a suitable model for the request. LISA Fiber, LISA Pro, and LISA Max map to the existing backend models, and each agent can keep its own choice.

AutoLISA FiberLISA ProLISA Max
LISA model selector with the four available options

Release status

Available now, and deliberately limited.

The mobile app is in active development. This page only describes what is present in the current build.

In this release

AI chat with conversation history; personal agents with roles, instructions, skills and models; agent groups; completed works; in-app notifications; approval workflows; model selection; light and dark themes; four interface languages.

Not currently offered

Sandbox and Store are temporarily removed from the mobile app interface. They are not advertised here as available features.

Distribution

There is no public download or APK release yet. A debug build used for internal testing is not a public release. Join the waitlist for availability updates.

See every capability.

Each feature on the features page is part of the real product, with a screenshot from the app.