Improve "Box" and "Dropbox" like integrations to allow search before attaching

This is related to the attachment field type. Our team can have hundreds of box files and folders, searching in the attachment integration here in airtable would leverage a lot of benefits for a variety of use cases for my team.

1 Like


Describe for me the search process. Would they be searching for references to documents that still exist in Box? Or are these documents now stored in Airtable?

Also, when you find something you’re looking for, do you want to see the reference of it from Airtable, or the original address of it in Box?

Read this document (Airborne Search) when you get a chance - it might help you describe your use case better.

Hi @Bill.French

I am unable to share an image or link so bear with me:

  1. click on attachment field (from grid or form)
  2. Select BOX integration
  3. SSO/Validate box login
  4. Populates my box folders than files in alphabetical order
  5. Attach file from box
  6. Done
1 Like

This is confusing me I think, but I kind of get it. Allow me to restate with a few more words. :wink:

The user must be able to browse through or enter data into an Airtable record and on an attachment field in the record, they must be able to pull documents from Box based on a search in Box.

Essentially you need to make it seamless to add select Box documents as attachments in Airtable from Airtable.

Is that correct?

And, of course, the inability to search in the current process makes it difficult to find the document you want before selecting it for addition into Airtable.

Okay - so here’s the rub - the current integration feature in Airtable is set in stone and there’s no easy way to extend the functionality.

There is a way to make this possible with the Airtable API, but I’m not certain that would be any more useful as it would be less seamless than adding it with the existing feature.

Hi Bill,

Not quite. Why don’t you create a box/dropbox account (should be a free trial period) and authenticate through Airtable. Then you will see screen/steps listed in #3-5 in my previous comment. Without that context, it’s going to be challenging for you.

You can in fact “browse” once you authenticate. You cannot sort, or search.

Nothing is “set in stone”! Thats the beauty of client feedback, continuous improvement, etc etc.

Thanks for your thoughts.

I saw all the steps - been a Box user since 2005 and I’ve had a number of delightful conversations with Aaron Levie, one of the founders.

It’s a great tool and your idea is a good one - findability is always a challenge and yet so important for productive work.

Indeed, and I certainly can’t speak for Airtable, but I was referring to the underlying integration that all the products supported by Airtable must fall under. Once an oAuth process has been completed giving Airtable access to your documents, the pathway is open and an API “interchange” begins. It seems that integration is intentionally homogenized - a lowest-common-denominator approach and not unlike all other SaaS systems that allow you to attract content in this way.

While the code to create the authentication process and enumeration of your documents is virtually identical for all document services (Box included), implementing search is likely far different for all of them as well.

Extending that functionality would probably require unique development efforts for each integrated system and also some vendor-specific development skills to create a search feature that would ostensibly be in (and for) a service that is outside of Airtable. I don’t see that happening especially when both vendors are likely to point to their respective APIs as the remedy.

Imagine the irony if Airtable announced an extended search feature for Box, but has nothing to say about search for its own content. :wink:

@Bill.French name dropping is irrelevant to the conversation, just comes across pretentious on the receiving end.

Your original response around “pulling” etc was what left me to imagine you might not be visualizing steps 3-5. If you did, then you understand theres no pull. it just populates.

oAuth offers many paths for enhancements, when my team was working on integrations this was a simple update via JS or Python.

Personally I’ve found airtable very invested in their product and enough people support an achievable enhancement its been resolved. We have a large enterprise use case so maybe our success manager is a better advocate then others :slight_smile:

I dont think this is a if/or feature request, which is why I submitted it. You can in fact tackle both and if one is a lower lift and enhances the integrations across multiple integrations including a google drive and drop box search etc who knows.

Have a great evening.

I suspected this was the case, and perhaps the Airtable team will take your suggestions to heart since this affects so many more users. I have upvoted it and anything they can do to streamline integration for non-technical users is typically a good investment.

Box also has a really superb search feature that drives adoption in large enterprises like yours. And it works pretty well when integrated into Airtable Blocks. This is how I search Box with one of my larger clients.

To integrate Box like this you have to use the URL they provide in their iFrame embed scripts because Box doesn’t allow its app to be framed into any other app (security thing I guess).

One very cool thing about their iFrame URLs is they also support deep linking and even filtering and sorting parameters. Anything you jam into the iFrame URL works in Airtable. As such, you can isolate areas of your Box assets in different Airtable dashboards and bases where the scoping of documents is more relevant to the tasks in specific bases/tables.

A URL that works in Airtable’s embed block looks something like this and fully support SSO.

In most cases - depending on browser security settings - once authenticated, you don’t actually have to give your Box account access to Airtable via oAuth. This may be of concern at some organizations, so this is a slight advantage - it lessens the security attack surfaces and management of shared accounts.

Box Search also works perfectly like this and when you’ve found a document you want to add to Airtable, simply copy the share link and drop it into the attachment field as a URL - it all works very nicely.

Indeed, it’s not perfect, but it’s dang close considering neither Box nor Airtable developers have ever had coffee. :wink:

Drag and Drop

I hang this final usability shortfall on Box, not Airtable.

Box implements the ability to drag a document from search results, but not outside its own HTML page boundary. As such, you cannot simply drag and drop from search results into Airtable attachment fields or vice-versa from Airtable to Box. This would be the most productive approach (I think). Box and Airtable working as a single app sharing documents seamlessly.

In my view, long before Airtable should attempt to build a replacement search feature into the integration layer, clearly a non-trivial task, Box should be encouraged to support drag events outside its app frame. Airtable already supports drop events.

Hi Bill,

How does the block/URL link solve for the attachment field type? Are you suggesting we abandon the attachment field type for a URL? We prefer the visualization from box to airtable for attachments. Very visual use case for our team.

Agree to disagree on what is and isn’t “trivial”. Neither of us manage this product so who knows what is and isn’t an incremental value, all we can do is submit the request.

I dont know any tool that does web to web drag and drop. Would not be interested either, as it requires a second screen open.

Good luck!

Not sure what you’re asking, but the desired use case described a search process was important when capturing documents into Airtable from Box. Since Box has a great search engine, I figure - why not leverage that? And when Box search is framed in a block it makes this possible without having a second web page or a tab or another app open. It seemed to fit your goal - I know it does for me and the client.

Not at all. Perhaps you’re not aware that attachments can be added by value or by reference (i.e., a URL). Since Box denies the ability to drag and drop (by value), the share link works well albeit with one additional copy/paste step.

This approach places Box search results immediately adjacent to the Airtable target location for your documents. It does this without running a separate app, it’s integrated, it leverages one of the best search engines in the document management segment, and it eliminates a security attack surface that most CIOs are nervous about.

Given the other approach, this seems like a good way to increase productivity and leverages the best of both Airtable (blocks) and Box (search).


Um, I don’t think so. If running Box fully in a Block (framed) it’s a single app - just Airtable - the screen shots show this, no?

Hi Bill,

I’m sorry if this got under your skin in anyway, this question wasn’t directed at you.

I’m concerned people will scroll through all this and misinterpret my original request and see your screenshots as a solve while I am still not able to.

Why would I hack a block and risk infosec cancelling my MSA altogether for this tool? We’ve had so many tools removed from our ecosystem already :frowning:

To be clear, I don’t want a separate block, window, etc to “search” box elsewhere. We solve that way right now. I want an improvement to that which is to integrate it into that solution your showing in your screen now. Everything else is a lateral move to what is already being done.

You’re right! This is beyond a discussion that elucidates your requirements. It should be an article for searching and pulling Box documents into Airtable.

I think it’s a mischaracterization to suggest the workaround I offered is a “hack” or has added security risks; it doesn’t. It uses a well-supported and an often-used block built by Airtable. iFrames are fully secure and based on open web standards. And, if anyone hacked an Airtable account, they would not have access to the framed Box account because it does not require an oAuth linkage.

Feel free to describe this as not a solution to your desired feature. I’m okay with that.

Once I understood your idea, I voted it up. I hope they add this feature.

I disagree, but if you want to share your suggestions on a different post for show and tell, that makes sense.

My use case is specific to the attachment field type integrations offered. Everything in blocks and iframe is outside of that field type and irrelevant to this question.

I know that seeing “box” made it tangentially related but mostly I feel like you had an idea before you heard or understood me.

Hi Bill,

Final note:

We tried your “share URL” as a URL attachment idea. Here is our feedback:

  1. going to box, searching box, clicking the share feature, waiting for the link to generate, copying the URL, inserting the link = 8 clicks. Whether this is in a “block” or not, I do not want my users to go through this process of sharing and copying URLs

  2. the URL is displaying a “login to authenticate” instead of displaying the attachment. What does that mean? That means that every single collaborator on this base also needs the provisioned access in box first for this to work, AND they have to login to box otherwise the “preview” attachment feature does not work. As I stated, we are an image driven/visual use case. The box authentication integrated into the attachment solution resolves all of these restrictions and issues I am describing.

Thanks for trying but next time, please listen.


I agree! I don’t want to do that either, ergo - why I upvoted your suggestion. :wink:

The process I use requires exactly 8 clicks to find anything through search, then select, and import a Box document. But [today], how many clicks are required to locate and import a Box document without search?

Assuming the document is in the root folder, it’s six clicks without the advantage of a search functionality. The reality is that it is likely going to be north of 15 to 18 clicks depending on the depth and complexity of your Box repository.

This is precisely why I agreed that your idea was a good one. It’s silly the way it presently works. The workaround I use is simply less silly - perhaps far less silly.

If you have dozens or hundreds of Airtable users who need to each install separate blocks to bring Box search into the process, that’s ugly too. Five users? Possibly not that ugly. But that’s an Airtable inefficiency; they need an enterprise admin and access control feature for commonly deployed block’s.

If, by company policy, you don’t like the approach I suggested, I understand. I also don’t like it - it’s not ideal, but it was intended to help you (maybe others) with a process where search was perceived (by me) as the most critical ingredient missing until such time as the feature can be upvoted and implemented.

I get it, you don’t like my observations. I’m sorry you feel that way. But, I think a forum discussion about new feature requests is supposed to include diverse ideas, workarounds, and alternatives. As such, I felt compelled to share mine with you and all the other readers who might benefit from a glimpse of at least one possible pathway that is not perfect, but better than no search at all in the process of introducing Box documents.