Questions historical societies and museums ask
Plain answers to the questions boards, archivists and volunteers ask before they send a first box.
These are the questions we expect from historical societies, museums and libraries weighing a first project. If yours is not here, write to hello@dastok.com, or to sales@dastok.com about a particular collection. For the method in full, see how we work and engagements.
What is a pilot box?
A pilot box is a small first batch of your real material, run through every stage: assessment, capture, text recognition, cataloging and return. You see the images, the recognized text, sample catalog records and the measured error rates before you commit a collection. It ends with a written report that includes the time each stage took, so the quote for the rest of the collection is built from measurements rather than guesses. The example collection shows one, with invented material.
How much does a project cost if you do not publish prices?
Cost depends on what is in the boxes: how many items and images, the formats, their condition, how much text needs correcting and how much description needs review. We quote each engagement in writing from its scope, and the quote states its basis (per image, per item or per hour) and what is excluded. A pilot box gives measured times for your own material. Until then, the digitization budget calculator lets you rough out labor hours with your own figures.
Do our originals have to leave the building?
No. Most collections can travel to our studio in Colorado Springs, packed, listed and receipted at every handover. For material that should not travel, because it is fragile, oversize, restricted or simply too much to move, we can work on site. Either way, the same capture specification and quality checks apply. Our collection care page explains handling, transport and chain of custody.
Can you read handwriting?
Yes. Handwritten text recognition works on letters, diaries, ledgers and registers, but how well it works depends on the hand, the ink and the paper. We measure it on a sample of your own pages before you decide anything, and we report each hand separately. Where recognition is weak, you can choose full correction by a person, correction of names, places, dates and numbers only, or uncorrected text labeled as such. The text recognition page explains the method.
How accurate is the text, and how will we know?
You get a measured number for each batch, not a promise. A person transcribes sample pages exactly, and we compare the machine’s text with them to get a character error rate and a word error rate for each kind of page. After correction, a fresh sample is measured again. The rates go into every batch report, and the OCR accuracy calculator shows how they are worked out.
Who owns the scans, text and records?
You receive the master files, access files, recognized text and catalog records. Ownership and licensing terms are set out in each agreement, which you read before any work starts. Your originals remain yours throughout, and the files are delivered in open formats so you can copy, move and use them without depending on us.
What file formats will we receive?
Typically TIFF preservation masters, JPEG access copies, page text as plain text and ALTO XML, searchable PDFs where useful, and catalog records as CSV plus Dublin Core, MARC 21 or EAD, whichever your system imports. Where publishing is in scope, you also receive IIIF manifests. Everything is packaged in BagIt bags with checksum manifests and an inventory. The exact list is written into the specification at the start of the project.
Do you host the collection, or can it go into our own system?
Either. We can build a search site with full-text and meaning-based search and IIIF image viewing, or prepare files for the system you already run, for example a CSV for Omeka’s CSV Import plugin, or EAD or MARCXML for ArchivesSpace. If we host anything, the agreement says how you leave with all your files and records. The collection search page has the detail.
Can you help with grant applications?
We can give you what most applications need: item counts, a capture specification, measured times and a budget, from a collection assessment or a pilot box. We cannot promise funding or eligibility, and programs change, so check each funder’s own page. As of November 2026, NEH’s Collections Stewardship program lists digitizing, cataloging and transcribing collections among eligible activities, with a deadline of December 15, 2026; IMLS’s FY2027 applications closed on November 13, 2026; and the NHPRC funding page lists no open round.
What about copyright? Can we put everything online?
Digitizing for preservation and publishing online are different questions. Most of a historical society’s box is unpublished: letters, diaries, ledgers and many photographs. In the United States, unpublished works generally stay in copyright for the author’s life plus 70 years, or 120 years from creation when the author’s death date is not known. Works first published in the United States before 1931 are in the public domain. The rights review worksheet helps you record a working assessment and a RightsStatements.org statement for each series, but neither the worksheet nor we give legal advice.
How long does a project take?
It depends on the number of images, the formats, the condition of the material, the correction level you choose and how quickly decisions come back at each approval point. Capture is often the longest stage, while text correction and cataloging can overlap with it. A pilot box measures how long each stage takes on your material, and the project timeline calculator turns those rates into a schedule you can plan around.
Will AI make things up in our records?
Models do make mistakes, including confident ones: a draft summary can mention an event the letter never describes. That is why every machine-suggested field is labeled until a person reviews it, and why we count how many suggestions were accepted, edited and rejected. Rights, restrictions and sensitive decisions are always made by people. Our AI policy explains the labels and what you can opt out of.
Will our material be used to train AI models?
Your images, text and records are not used to train any third party’s general-purpose model without your written permission. If a recognition model is trained for your collection, the agreement says who may use it afterward, and the pilot report names every engine, model and service used on your material.
Do we need to sort or inventory everything first?
No. We work from what you have, and a rough list or an old accession register is enough to start. A collection assessment builds the inventory with you. If you want a head start, the collection inventory worksheet has the columns we use.
Do you work outside Colorado Springs?
Yes. Our studio is in Colorado Springs, and we work with collections across Colorado and beyond. Originals can come to the studio by an agreed method, or we can work on site. Text recognition, cataloging, search and preservation work can also start from images you already have, wherever they are. Our Colorado Springs page has more.
Send one box first.
A pilot box shows you the images, the recognized text and the catalog records before you commit a whole collection. Tell us what is in the box and what you want people to find.
Studio and contacts
24 South Weber Street, Suite 225Colorado Springs, CO 80903(719) 244-8450sales@dastok.com