iMpact — AI work automation software and the ORION assistant. Technologythat matters.
iMpact builds software and intelligent systems for real work — designed with restraint, held to a high standard, and always under your control.
In 10 seconds
What it is
An AI assistant that plans your computer tasks, step by step.
Why trust it
It shows every step and asks before it acts.
Where we are
The website version is free today. The desktop app is in progress.
- ORION
- In development
- Labs
- Exploring
- Early access
- Design target
Summarize this thread and draft a reply I can edit.
- 01Understand
- 02Plan
- 03Choose tools
- 04Act
What we believe
Four commitments we design against.
- 01
Useful before impressive
We start from real work and real constraints. If a feature does not make a task shorter, clearer, or calmer, it does not ship.
- 02
The person stays in command
Software that acts on your behalf should show what it is doing, ask at the right moment, and stop the instant you say so.
- 03
Restraint is a feature
Fewer surfaces, clearer language, quieter interfaces. Depth is available when you want it, never demanded up front.
- 04
Built to last
We choose durable foundations over quick wins, and we would rather name what is unfinished than imply it is done.
First product
ORION. Personal intelligence, reimagined.
Our flagship product in development: a computer assistant designed to understand permitted context, work within limits you set, and show each step it proposes.
Explore ORION- 01
Visual context
We are developing a permission-bound way to interpret the interface and relevant task context.
- 02
Controlled action
The product is being designed around visible steps, explicit permissions, interruptibility, and reversibility where possible.
- 03
Model routing
A provider-agnostic routing foundation is in development around capability, latency, privacy, and cost.
- 04
Accessible paths
Keyboard, voice, and assistive input are design goals being considered alongside the visual interface.
How far along is ORION? Here's what's built and what's left.
ORION build progress
43%
built · 57% to go
- Website workspace (plan + confirmed actions)100%
- Task planning layer80%
- Visual context engine45%
- Permission-aware action layer35%
How it works
From understanding to action, one visible path.
Select a stage to explore the accountable workflow we are building for ORION.
01 / 05
Understand
The intended flow reads the window, selection, and context you allow, then restates the task so it can be corrected before anything runs.
Labs
ORION today. More deliberately, later.
We would rather ship one product properly than announce five. Future work is listed only once it is genuinely being explored.
- Exploring
Model routing research
Measuring how far requests can be matched to smaller or on-device models without a loss in usefulness. Internal research, no product yet.
- Exploring
Ambient interface studies
Prototypes for surfacing assistance without a chat window: keyboard-first flows, quiet notifications, and voice as an equal path.
- Design target
Open slot
Deliberately empty. When a second product is real enough to describe honestly, it will appear here — not before.
Human control
Capable software should still answer to you.
- Local first where it counts
- Sensitive work should stay on your machine. Where a request must leave the device, ORION says so before it happens.
- Permission, not assumption
- Access to files, apps, and the network is granted per capability and revocable at any time, without hunting through settings.
- No silent training
- Your content is not used to train models. There is no hidden opt-in and no quiet change of terms.
- Accessible by design
- Keyboard, screen-reader, and voice paths are built alongside the visual interface, not retrofitted after launch.
Where this goes
A long-term company, built in the open about what is finished.
Nothing on this page describes a shipped, generally available feature.
- Visual context engine
- Permission-aware action layer
- Multi-provider routing foundation
- Keyboard and voice workflows
- Readable activity history
- Granular privacy controls
- How far on-device models can carry daily work
- Personal workflow memory without surveillance
- A safe extension interface for developers
Early access
Follow the work from the beginning.
Join the early-access list. You will hear from us when there is something real to use — nothing else.