Independent research. Real machines.

Hardware,in the open.

What works. What breaks. What changes.
We test hardware in Santiago and publish the evidence for everyone.

SANTIAGO, CHILE / NODE 001
FIG. 001 / INSIDE THE MACHINE
Exploded view of a computer assemblyA custom technical illustration separating the heatsink, circuit board and chassis to reveal what is inside. This is a concept drawing, not a tested lab unit.01 / THERMALS02 / COMPUTE03 / BUILD
CONCEPT ASSEMBLY · NOT A TESTED UNITPLATE 01

Hardware intake open

Building the public inventory.

Public units
00
Active runs
00
Published reports
00
View the register
01 / THE PRINCIPLE

A machine is a
starting point.
Evidence is the work.

What “open” means here

The process is open: methodology, provenance, disclosures, evidence, expenses and upstream work. We don’t claim that every computer we test is open hardware.

One machine can answer many questions, across operating systems and years. Test locally. Publish globally.

Read our independence policy
02 / RESEARCH DIRECTIONSFOUR OPEN TRACKS

Better questions.
Useful answers.

Compatibility goes beyond a spec sheet.
These are the questions shaping our queue.

01 / Omarchy compatibilityIN THE QUEUE

A clean install is only the beginning.

What happens when a real machine becomes a daily driver? Follow installation, suspend, graphics, media and updates through repeatable runs.

  • Installation & boot
  • Suspend & resume
  • Graphics & displays
  • Media & connectivity
02 / Retail Linux hardwareIN THE QUEUE

Same model. Different machine?

A model name rarely tells the whole story. Examine exact retail configurations, regional SKUs and the components people can actually buy.

  • Exact configuration
  • Regional components
  • Input & connectivity
  • Firmware & kernel
03 / Agent workstationsIN THE QUEUE

Does it hold up when the work gets real?

Investigate containers, hosted-agent workflows, local inference where practical, and sustained developer use with bounded, documented workloads.

  • Containers & virtualization
  • Sustained thermals
  • Power behavior
  • Developer workloads
04 / Regression watchIN THE QUEUE

It worked yesterday. What changed?

Keep the hardware in the loop. Repeat runs after firmware, kernel and operating-system updates to make improvements and regressions visible.

  • Firmware changes
  • Kernel updates
  • Repeated test runs
  • Upstream issue history

TRACKS ARE RESEARCH DIRECTIONS. THEY DO NOT GUARANTEE A TEST, SCHEDULE OR PUBLICATION.

03 / THE METHODREAD THE METHODOLOGY

Every result
has a paper trail.

Units, runs, reports, retests.
A record you can follow. A method you can repeat.

  1. 01
    UNIT

    Know the machine.

    Register an OHL ID, exact configuration, regional SKU, source, ownership and provider relationship.

  2. 02
    RUN

    Ask a useful question.

    Choose an operating system and a bounded test. Record the method, versions and conditions.

  3. 03
    REPORT

    Show the evidence.

    Publish privacy-scrubbed observations, failures, workarounds, expenses and upstream links.

  4. 04
    RETEST

    Keep asking.

    Run it again after firmware, kernel and operating-system changes. A result is a record, not a lifetime verdict.

METHOD STATUS

Baseline categories and evidence rules are published. Executable test definitions will be added before the first public run.

04 / THE PUBLIC RECORD

Nothing hidden.
Nothing invented.

The lab starts empty, not imaginary. No physical unit has been accepted into the public register yet.

Hardware intake is open. The method is taking shape. This is where the evidence will go.

Follow the work on GitHub
OHL / REGISTER INDEXPUBLIC
Current Open Hardware Lab records and status
RecordStateSource
Hardware inventoryProvenance & custodyAwaiting units
Baseline methodologyTest categories & evidence rulesPublished
Reports & retestsResults, failures & upstream workAwaiting runs
Lab ledgerSupport, expenses & allocationsPolicy published

CURRENT PUBLISHED STATE · MAINTAINED IN THE REPOSITORY

05 / MAKE THE WORK POSSIBLE

Put hardware
to public use.

Have a machine worth investigating?
Help turn it into a reproducible record.

Hardware pool

There’s more than
one way in.

Owned machines, approved loans, permanent placements and retail purchases enter the same public inventory.

We select work from available hardware, funds, research value and lab capacity. Campaigns are optional procurement tools.

Offer hardware
01

Temporary loan

A bounded research loan with arrival condition, disclosed ownership, return date and prepaid return logistics.

02

Permanent placement

Ownership transfers to Arca Computer, Inc. The machine stays in the inventory for long-term regression testing.

03

Retail delivery

Fund or supply an exact unit through an approved retailer or distributor, with warranty and customs responsibilities resolved.

Before anything ships

Get written intake approval first. Send the model, exact configuration, location, ownership proposal and shipping plan to ohl@arca.computer. International intake must resolve shipping, customs, warranty and return responsibilities; local distribution or DDP is preferred.

The unit must have clear research value, workable logistics and a publicly disclosed relationship. No favorable-result or editorial-control conditions are accepted. A suggestion does not guarantee an operating system, schedule, report or positive result.

Lab fund / voluntary support

Fund the work.
Free the findings.

Help cover equipment, testing and public evidence.
General support stays in the unallocated OHL pool until a specific use is published.

Follow the money
ONE-TIME SUPPORT · USD
Your call

You choose the amount at checkout.

Support with Stripe Read the support terms

Voluntary support to Arca Computer, Inc. Not a charitable contribution or tax-deductible. No goods, services, reserved tests or promised results.

Our independence isn’t for sale. No paid verdicts. No editorial approval. No removal of legitimate failures. Support buys lab capacity, never a conclusion.