Contract engineering · GeneralQuery Technologies · Berhampore, India

We build the system. Then we make it able to tell a person is there.

QueryLabs is General Query's engineering practice. We take contract work: small team, senior hands, production systems. We're also the only team that deploys QueKey, the identity layer our lab builds.

Capacity · Q3 2026 Updated 03 Aug 2026
Committed 2 engagements
Reserved · lab QueKey · Strata
Open 1 engagement
Committed Lab Open We publish this because the lab has first claim on the calendar, and you deserve to know what's left before you write to us.

The practice

No bench. No account managers. The people who build QueKey are the people on your project.

Shape Team size · small, senior
Model · contract, defined term
Concurrency · three engagements, maximum
Working hours · embedded with your team
Reporting · directly to you

Most agencies sell you a delivery org and staff it with whoever is free. We can't do that and don't want to. QueryLabs is a working practice attached to a research lab, which means the constraint is real: we take a small number of engagements at a time because the rest of the week belongs to QueKey and Strata.

What you get in exchange is the thing agencies charge extra for and rarely deliver. The engineers who designed a patent-pending authentication framework are the engineers writing your migration. There is no layer between the brief and the build.

We work best when we can see the whole system: the product, the data model, the internal tools, the places where humans are still doing what software should. Scoped narrowly we're still useful. Scoped widely we're worth the constraint.

Fewer clients is not a limitation we apologise for. It's the reason the work is good.

Engagements

Three ways this starts. Most of them start in the middle.

You do not need to know which one you need. Describe the problem and we'll tell you which of these it is, including when the answer is that you shouldn't hire anyone yet.

What QueKey actually is

A credential proves the account. It has stopped proving the person.

QueKey Type · OAuth 2.0 identity provider
Asserts · human presence
Protocol · IAF, patent pending
Integration · days, not a standards migration
Deployment partner · QueryLabs, exclusively

Forging a person used to be expensive. A face, a voice, a document, a plausible history: all of it is cheap to synthesise now, and convincing enough. Okta, Auth0, Clerk and WorkOS verify credentials. Passkeys verify a device. Every one of them answers is this the right account.

QueKey answers the other question. It's an identity provider your application consumes over ordinary OAuth, and what it returns is an assertion that a human being was present at the moment of the request.

That's useful the first time at login. It's more useful everywhere else: onboarding you need to trust, a research panel that shouldn't contain synthetic participants, a payout above a threshold, a session that's been open for six hours and can no longer vouch for who's driving it.

QueryLabs is where that gets deployed. General Query builds the layer. We put it inside real products, and what breaks in the field goes back to the lab.

What we work on

The substrate. Unglamorous, and the reason anything else is possible.

Product and platform

Web applications and SaaS products built from scratch or taken over mid-flight. Architecture that survives the next two years of your roadmap.

Most engagements

Integration

Making systems that were never designed to talk to each other do it reliably. Legacy stacks, vendor APIs, data that lives in four places and agrees in none.

Where teams get stuck

Workflow automation

The manual steps between your tools. CRM and messaging orchestration, scheduled pipelines, the operational glue nobody wants to own.

Fast payback

Interface design

Screens designed against real use, not a component library. We design what we build, which keeps the handoff honest.

Included, not upsold

Brand and identity

Marks, type systems and the written rules that keep them intact after we hand over. Guidelines you can hold someone to, not a mood board.

Published as a book

Identity and access

Auth that holds up: session handling, permissions, onboarding flows, and presence verification where the stakes justify it.

Our specialism

Applied AI

LLM integration inside real products, model training and fine-tuning, and generative AI features scoped to what your data can actually support.

Growing fast

Context systems

Strata, the lab's research into structured memory across AI assistants. Not a service yet. Say so if you want to hear when it is.

In research
We work across TypeScript React Node Python Postgres MongoDB AWS Google Cloud Docker OpenAI & Claude APIs

The record

What we've shipped, and what we haven't yet.

This page will name clients when those clients agree to be named. Until then it lists what exists and what stage it's at. We'd rather be briefly boring than quietly inventing a portfolio.

01
QueKey

Identity provider asserting human presence. Built in-house at General Query, deployed by this practice.

Live
02
IAF

Interoperable Authentication Framework. The protocol underneath QueKey's verification model.

Patent pending
03
Client engagements

Contract builds in delivery and two published case studies. Write-ups go up as each one closes and permission comes through.

In progress
04
Strata

Structured memory across AI assistants. Research stage at the lab, no client deployments.

In research

Why a lab has a practice

The engagements pay for the research. They also correct it.

Structure General Query · the lab
QueKey · identity, live
IAF · protocol, patent pending
Strata · context, in research
QueryLabs · this practice

GeneralQuery Technologies is a research company working on two questions: whether software can tell that a person is present, and whether a person can tell what a system understands. QueryLabs is the commercial arm that funds that work.

That's the honest version and it's also the useful one. A liveness engine fails in a lab the way you designed it to. In the field it fails differently: bad lighting, a five-year-old front camera, a shared device, a user who won't move their head. We only find those cases inside real deployments, which means the contract work isn't a detour from the research. It's the instrument.

So when you hire us you're getting a team with a reason to care how your system behaves eighteen months from now. That's not a values statement. It's an incentive.

Read the lab's thesis →

Send a brief

Tell us what's broken. We'll tell you honestly whether we're the right team.

No discovery call to schedule a discovery call. Write to us in whatever detail you have, and you'll get a real answer from someone who would be doing the work.

What we'll ask

What the system does today, and where it stops doing it.

Then

Who it's for, and what changes for them if this works.

Then

What's already been tried, including what failed.

Finally

When you need it, and what you've budgeted.