Bring your organisation's collection into OHM
Available now
For organisations that already have a body of open hardware designs — a university's project archive, an NGO's field equipment documentation, a programme's accumulated output — and want it discoverable rather than entered by hand.
If you have five designs, add them individually. This page is for the case where you have two hundred.
The realistic picture
There is no single import button, and honestly there can't be one, because collections don't arrive in a single format. What actually happens is one of three things, in descending order of how easy your life will be.
Your data is already in a format OHM reads. If your designs are published as OKH-LOSH manifests, conversion is a solved problem — there's a converter, and it runs over a batch without anyone writing new code.
Your data is structured but in a format of your own. This is the common case for organisations that documented things properly before any standard existed. Your spreadsheet or database is perfectly good; it just isn't a format anything else speaks. Here we typically write an adapter for your format specifically. That has been done before — the datasheet converter in OHM exists because Médecins Sans Frontières documents its open hardware work in a particular internal way, and meeting that where it was proved easier than asking an organisation to redocument years of work.
Your data is unstructured. Documentation in prose, PDFs, or scattered wiki pages. Extraction can get part of the way, but expect review rather than a clean import.
Why we write adapters rather than asking you to convert
It would be simpler for us to publish a format and ask everyone to produce it. It also wouldn't work — see meet people where they already are.
An organisation with two hundred documented designs has already done the hard part. Asking them to redo it in our shape, before they've seen any benefit, is how a project ends up with excellent tooling and no users.
What building an adapter involves
Concretely, if your format needs one:
What we'd need from you
- A representative sample — a dozen records covering the messy cases, not the tidiest ones
- What the fields mean, particularly anything domain-specific. Field names are rarely self-explanatory across organisations
- Which fields matter to you. Not everything needs to map, and pretending otherwise slows things down
- Someone who can answer questions about the data as they arise
What you'd get
- A converter from your format to OKH, runnable repeatedly, so this is an ongoing pipeline rather than a one-off migration
- Your designs in the catalogue, matchable against facilities
- The converter itself, in the open, so anyone with a similar format benefits
What to expect around ambiguity. Real collections always contain records that don't quite fit — missing versions, ambiguous licences, files referenced but never published. We'd rather flag those for review than silently guess, so expect a list of things needing a human decision. That list is a feature.
Doing it yourself
If you'd rather not involve us: conversion and creation are available through the API and the command line, so a technically comfortable team can write their own adapter against the same interfaces we would.
The CLI is usually the better fit for bulk work — see use the OHM API.
There's no bulk import in the web interface yet
Batch import currently runs through the API, the command line, and purpose-built scripts. A web interface for it is planned and would make this considerably more approachable. See what's built and what isn't.
Before you start
Two decisions worth making deliberately.
Where it goes. If your organisation has any constraint about where its data lives, decide that before importing rather than after. Running your own instance means the collection is yours end to end — see run your own node.
Who sees it. Imported records follow the same visibility rules as anything else, and default to private. You can bring a collection in, look at it, and decide about sharing afterwards. See who can see your data.
Getting in touch
If you have a collection and want to work out which of the three cases you're in, get in touch. The answer is usually apparent from a handful of sample records.