# Labs · Daypart

> We don't take the vendor's word for it. We measure it. A live wall of real devices and a materials bench, feeding signal straight into the Research OS.

Source: https://daypart.ai/labs

Daypart Labs

## We don't take the vendor's word for it. We measure it.

Daypart Labs is our first-party collection and testing operation: a live wall of real devices and a materials bench. Signal and results feed straight into the Research OS, so the deepest diligence questions get answered with evidence we generated ourselves.

01 · Device testing lab REAL HARDWARE · NOT EMULATORS

### A wall of real devices, watched and driven.

Connected TVs, iOS and Android phones and tablets, streaming sticks, smart speakers, wearables, and connected-home hardware. Some sit collecting continuously; others get pointed at a specific question and run under controlled conditions.

AMBIENT COLLECTION

Devices run around the clock, capturing what real hardware actually sees: ad load, curation decisions, delivery by geography, app and SDK behavior, network traffic.

CONTROLLED TESTING

For a specific matter we script the scenario: swap identities, change geographies, drive the app, and record exactly what happens, replayable and hashed.

ALWAYS EXPANDING

We add new collection surfaces constantly, both to widen what the knowledge graph observes ambiently and to stand up whatever a client matter demands.

CONNECTED TV

ad load · curation · geo

iOS DEVICES

app behavior · SDK

ANDROID

install flows · perms

TABLETS

rendering · UX

SMART SPEAKERS

voice · audio ads

WEARABLES / IoT

connected signal

[DEVICE COUNT] DEVICES ON THE WALL ● CAPTURING

02 · Materials & sample bench WHAT IT IS, NOT WHAT THE LABEL SAYS

### Physical evidence, held to the same standard.

When a claim rests on what a product is made of or how it performs, we test the thing itself. Mass spectrometry, teardowns, and sample assays turn a marketing claim into a measurement, and the result attaches to the entity in the knowledge graph like any other artifact.

MASS SPECTROMETRY

Composition and contamination against a reference.

DEVICE TEARDOWN

What is actually inside the hardware, part by part.

SAMPLE ASSAY

Does the sample perform the way it is represented?

CHAIN OF CUSTODY

Sourcing to result, logged and defensible.

MASS SPECTRUM · SAMPLE 04 REFERENCE MATCH 0.97

SYNTHETIC DEMONSTRATION · SPECTRUM SHOWN FOR ILLUSTRATION

Into the Research OS

### Lab output is evidence, so it carries provenance from the first capture.

Nothing from the lab is a loose screenshot or a number in a slide. Each capture and each result enters the OS hashed, timestamped, and tied to the device or sample that produced it, ready to be traced back when a finding is challenged.

### Have a question only first-party evidence can answer?

Tell us the claim you need tested and the decision it informs.

[Start the conversation →](https://daypart.ai/contact)