Calling all photo buyers: Are you sick of having a gazillion tabs on your desktop while toggling through different photo agencies for the stock image you need? Former  Vanity Fair and Travel + Leisure contributing photographer Andrew Rowat and technology developer Vache Asatryan created Haystack— which launched its beta in October– because they believed the previous technology was “like searching for a flight in 1999.” Here’s why they’re banking on what has already been billed as the Google Image search for stock photo agencies.

What was the inspiration for Haystack?

Andrew Rowat, CEO of Haystack
Andrew Rowat, CEO of Haystack

AR: Vache and I were on a shoot for an Italian design magazine. It was pretty full on—dozens of locations in a week. Putting together the call sheets each day was one of the biggest challenges. While scouting photos for each location, we thought there must be a type thing where you can get a ton of different photos in a single stream and be able to pick and choose. There were things like Flickr and Panoramio and Google, but what we were looking for didn’t exist. I reached out to several people and basically opening 20 tabs in your web browser was the only way it could be done. I think it is astonishing that in today’s day and age of technology that the process of licensing is as painful as it is.

How does Haystack work?

AR: Haystack is a search engine that just happens to have photos, it searches for images among our partner stock agencies. We currently have 19 partners live on the platform with a whack more on the way.  We’re taking all the metadata that we’re getting from our partners and we’re normalizing it. For example, one agency may have this field as “titles” when actually it should be “headline.” We normalize all that, and then layer on an additional layer of artificial intelligence machine generated tags. So when you do a search on Haystack, you frequently find results on our site that don’t surface up on the native stock agency site.

VA: We’re using different search technology, we’re taking into account more facts, more signals, more data points when it comes to the image. It’s really hard to aggregate data together, it’s really hard to normalize that data. It’s really hard to do a search engine. And to add image processing to the mix, it’s really hard job to do that too.

Haystack image search in action.
Haystack image search in action.

Are you a new kind of photo agency?

AR: No, we’re not competing with our photo agency partners.

VA: You can’t replace editorial curation. A lot of companies say that it can all  be replaced with AI. But there is a reason why agencies exist and that is to work with the photographers. Photographers don’t always know their best photo, but the agency knows the trends and what buyers want. You can’t replace that. We get asked all the time whether we’re going to work directly with the photographers, but we don’t have the curating skill to bring this kind of content together. We want to support out photo agency partners, keeping them independent and strong so they can continue their art of curation.

Where do you see Haystack a year from now?

AR: We want to narrow the gap between how you’re accustomed to searching on Google and to how you’re accustomed to searching on a stock site—which are very different things. How people search on a stock site is not typically how you search on Google or even just natural language. So we want to be able to bridge that gap or narrow it so you can just type out what you’re thinking and get exactly what you’re looking for.

VA: We’re also working on creating functionality within Haystack to help our users and our partners to save them time and make working collaboratively with their team members more seamless. We’re looking at WordPress, Medium, Twitter, Tumblr, etc and exploring Haystack integration into those platforms. Currently, users have to get photos from outside those publishing platforms and upload it. We believe there is an opportunity for tighter integration using Haystack.

Whose idea was it to include jokes while you wait for the search results?

AR: You can thank/blame me for the jokes.  We may yet expand to knock-knock jokes!


Mirrored Haystacks” by  Olivier Degabriele under CC BY-NC-ND 2.0, photo cropped from original.

Author: Connie Yu

Connie Yu has over 17 years in digital media and publishing. Most recently she was Photo Director of The Intercept, where she oversaw the creative direction and operations for the photography of the website. Previously she worked at AOL and

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.