sondalab

Sondalab · a lab for operators

Wired for operators,not passengers.

see it · steer it · own it

See what we wireWhat we believe →

The stance

Tools that keep you in command of the machines working for you.

Sondalab builds software that runs, watches and improves the agents doing your work, and keeps them legible, controllable and yours. The stance holds as machines grow more capable: from today's software agents toward the physical systems, and robotics, ahead.

Operator's log · SPEXR

The whole floor, every session, glows on one dark wall where you can see it breathe.

SPEXR, the agents view: four agent sessions on one project. Each card shows what the session is doing now. One card is marked 'needs you'.

SPEXR · the agents viewOpen the product →

  1. See it

    Every session says what it is doing now, in plain words, while it runs. No black boxes.

  2. Steer it

    A session that needs you stops and says so. You direct it, correct it or take the controls back. Assistance, never autopilot.

  3. Own it

    Your project, your harness, your machine. The data and the off-switch stay yours.

Five tools · two experiments

Many tools, one throughline.

Together they cover the whole loop of working with agents: writing and running the work, watching every step, scoping what a session loads, and hosting it on hardware you own.

  1. SPEXR

    The spec-based IDE where agents do the work and you keep command.

    IDEPublic beta

  2. introspect

    Audit and observe how your coding agent runs, replay and live.

    AppBeta

  3. claude-loadout

    Start each session with only the tools it needs.

    CLI

  4. local-agents

    Your agents, on your machine.

    App

  5. skillkit

    A pick-and-install kit of agent skills.

    Kit

  6. ads

    An agent team tests, changes and reviews each change. You keep the last word.

    CLIBeta

Research

Experiments, not products. Each one tests an idea and reports what it found. DAOS is finished and archived. MxM is active: its solution is being polished for production.

02

  1. daos

    A research bench for getting the same answer twice.

    LibraryArchived

  2. MxM

    Your coding agent remembers every session, on your machine, where you can read it.

    CLIActive research

The method

We run on our own tools first.

Sondalab is built by directing agents with the same tools we ship. What reaches you has already carried our own work.

  1. We write the intent first.

    Every change starts as a written decision, before an agent touches a file.

  2. The machines show their work.

    Each step is visible while it runs and traceable afterwards, so nothing lands unexplained.

  3. Drift doesn't ship.

    A design audit and a content schema stop the build on any mismatch. This page passed both.

A spec open in SPEXR: five acceptance criteria as tasks, two ticked, and the seven workflow steps from Specify to Ship with Implement in progress.
This site's own kind of work, in SPEXR: the spec first, its acceptance criteria ticked as the agent lands them.

Most software asks you to trust a black box across the wire.

We'd rather hand you the lid and a flashlight.

Contact

Building something you'd rather run yourself?

Sondalab is early and independent. If the operator's stance resonates, as a user, a builder or a collaborator, we'd like to hear from you.