We build our own company on this.
These are systems we build, run and release ourselves. Each one says what it does, what state it is in, and what it shows we can build for you.
About client work
Client work is shown only with the client's written permission. Until we have that, this page shows systems we build, run and release ourselves.
Featured systems
Each one starts from the problem it was built to solve, then says what we engineered and what state it is in.
Daena
System we operateAgent frameworks make it easy to let a model act. They make it hard to know what it did, why, on whose authority, and to stop it before an irreversible step.
A governed multi-agent platform: AI departments that plan and act inside approval gates and an audit trail.
ragX
System we operateA retrieval system that always answers will sometimes answer from nothing. For internal knowledge, a confident wrong answer costs more than no answer.
Our retrieval engine: answers from our own documents with citations, and a verifier that makes it abstain when the sources do not support a claim.
MCP Switchboard
Open sourceConnecting AI assistants to business tools means one configuration per tool per assistant, secrets scattered in config files, and no record of what the assistant did.
One governed endpoint that brings many MCP tool servers to Claude and ChatGPT, with an encrypted vault, per-tool policy, approvals and an audit log.
Our product websites
Website we shippedA technical offer is hard to judge from a page of claims. A buyer who is not an engineer needs to see the work happen, and a site that only describes it leaves the first step to email.
mas-ai.co, daena.mas-ai.co and kya.mas-ai.co: designed, written, built and deployed in-house.
Also built by MAS-AI
The rest of what we have built, with the label that says what kind of work it is.
AI Company OS
System we operateThe operating layer we run MAS-AI on: several AI coding agents sharing one memory, computed routing and deterministic checks.
MergeLoop
Open sourceA model council: one task sent to several AI workers, one synthesized answer back.
Lab
These are experiments. They are research in progress, not finished products, and each entry says how far along it is.
KYA Mission Control
R&DA trust layer for AI agents on missions: passports, bounded authority, budget stops and signed receipts anyone can verify offline.
Where it stands Pre-alpha research. The Mission Lab demo is public; the code is not yet released.
What the labels mean
System we operate
Software we built and use ourselves. Each entry says what state it is in.
Open source
Code we have released publicly on GitHub, so you can read it before you trust it.
R&D
Research still in progress. It is not a finished product, and the entry says how far along it is.
Website we shipped
Sites we designed, wrote, built and deployed ourselves, for our own products.
Bring us the bottleneck.
Tell us what is slow. We will say whether a build, a Blueprint or nothing at all is the right next step.
Bring us a bottleneck