"Transparency is a team sport."

Dillon Bergin is a data reporter with MuckRock, a nonprofit news outlet and support organization. He currently directs the Data Liberation Project. This interview was conducted by Morgan Kriesel on June 24, 2026.

“…he wanted to start a project that did that from the very beginning—made data public just because it's a public good…”

I wanted to start by getting into your career a little bit. How did you get involved with MuckRock? 

I remember first using MuckRock as a local reporter. I was a Report for America corps member with an organization called Searchlight New Mexico, and I remember filing some of my first public records requests—IPRA (Inspection of Public Records Act) requests, as they call them in New Mexico—and I had questions about the language I should use, or who I should file to, and somehow stumbled upon MuckRock. And I started using the MuckRock repository of all requests that other users have filed to shape my own request.

At that time, I was reporting on evictions and I wanted to know how to file some requests related to evictions. So I typed in the word “eviction” in the MuckRock search bar, and I was like, "Oh my gosh, I can see what people have filed in different states!" And then I copied their requests. I didn't have a MuckRock account at that time, so I was just copying and pasting them into my own email—which is totally fine [for other investigators to do as well], that's what all those public requests on MuckRock are there for—and that's how I first engaged with MuckRock.

I think I did eventually ask my editor to pay for an account, or maybe I even paid for it myself. [The repository of others’ requests] is free. The part that costs money is filing your own requests through MuckRock. The reason it costs money is we have a support staff that basically helps you through the requests at different points.

So, if you're filing to the smallest county clerk in West Virginia, and they want you to file the request through snail mail—sending an actual form to them, and they'll send you back a USB drive—our support staff will help send that mail. Then that'll get sent to an office, and our support staff will sort your mail, then make sure it's uploaded into your MuckRock account, and then you'll get it.

They help with bigger questions as well. Questions like, “Should I appeal this request?” Or, “The agency is saying that I can't file this request because I'm not a citizen of the state, and I don't have a Tennessee ID,” for example. So they help a lot with the logistical sides of filing the requests.

How did you become a part of the MuckRock team?

I started working on a project, a grant-funded project that was also partially funded and supported by MuckRock, called Documenting COVID-19.

I was just working on a contract basis with them. Then maybe a year-and-a-half into that project, they were like, "This is working well, and we wanna take this team and make it a part of MuckRock." And that eventually grew into the editorial team at MuckRock.

So you helped shape the editorial team?

Yeah. There were maybe different iterations before me, but at that point, they were kind of rebuilding it.

I know the Data Liberation Project (DLP) was originally an independent project that was adopted by MuckRock in 2024, is that when you took over as director?

Yeah, exactly. Before, it was run by Jeremy Singer-Vine. I can only tell my recounting of his origin story, which I was sitting, eating a doughnut and drinking coffee [while Singer-Vine told it].

In my understanding, he had started the Data Liberation Project after going on a hike. While he was on the hike, he realized that a lot of what he does as an investigative data journalist usually happens at the end of a long investigation. At least in the newsrooms he was in, after one or two years (or however long the investigation took) you make the data public. If you do. A lot of newsrooms don't, unfortunately. But he was thinking, "As soon as I get that data, it's already useful to the public. And it's supposed to be for the public.”

His moment on the hike, as I remember it, was realizing he wanted to start a project that did that from the very beginning—made data public just because it's a public good, and not because he was reporting on a story, and publishing it after.

Right. So enabling more journalism, rather than gatekeeping your sources.

Yeah. Or maybe, and I think this is where he was going, it is important as well for advocates, for attorneys, for civic groups, for interested people—that data could be really useful to them, sooner rather than later. So I think both for journalists, and all these other groups, making the data public first could be more impactful than waiting longer.

I know you guys [preserved] the EPA Risk Management Program data, which was taken down January 2025. But I can't remember if you guys have done any of the other big data sets that have been taken down or restricted. Are you working on any of those?

No, I think that's the biggest one that we've been able to contribute to, and that's something, again, that predates me.

But I know that EPA one is really important, we have heard from lots of people about it. I was at a journalism conference and met someone from the Union of Concerned Scientists and she was like, "Oh, I’ve been wanting you to talk to you about this dataset that you have up on the website. [The DLP site] has been the one way we've been able to access it.”

Yeah, that was huge. I just did a piece about the recent chemical explosions on the West Coast, and the Risk Management Program data is super tied into that. It was insane that this dataset went down [because of] chemical lobbyists’ requests to the administration. It’s really dangerous, and a huge story. So the fact that we did have this dataset because of your guys' work is really huge.

Every alternative source [for this removed data] that I found—the Coalition to Prevent Chemical Disasters (which is manned by the Union of Concerned Scientists), the Environmental Defense Fund, and Drexel University—they were all using the data that you guys preserved, and they all cited you.

That's awesome. And that's part of the ethos of DLP. I don't know if Jeremy or the volunteer that worked on it several years ago could have imagined where we are in 2026, and how useful this data would be. They were just trying to get the data, and get it into a shape that people would be able to explore.

That must be exciting. These things really do make a difference to research and to advocates. So, it's really important, and I'm glad you guys are doing it.

Thanks.

“We want to find… people in the public records world, and give them tools and community—things to make [this work] not disillusioning and lonely. ”

[Before] you took over, it seems like FOIA requests were the main way data was being “liberated,” and some web scraping was going on. Did that process change when you took over, and especially after Trump's second term started? Are FOIAs harder to get answered these days?

Generally yes, from everything I've heard, and experienced myself.

That is not really what changed our approach, but rather, taking Jeremy's idea of data could just be made public because it's a public good—this creates a very open and transparent process that would be really useful for volunteers to be involved in. I think there were already volunteers helping Jeremy, but he was kind of doing it on a one-to-one basis I believe. Like, someone has an idea for a request, they collaborate on a request, they send off the request.

At MuckRock, we wanted to orient it more towards being a volunteer-driven project as well. And focused not just on getting the requests off and making the data public, but building it as something really participatory. So that's been the reshaping, or really the slight refocusing of DLP, which is to do all these things but do it while also encouraging anyone—whether you're a data journalist or not—to participate in finding datasets that the public should know about.

So instead of specific individuals requesting [specific datasets], it's more like you're all working together to select datasets?

Yeah. We have a meeting and people throw out ideas for requests that they think are useful. The way that Jeremy had done it, and it's still happening, is organizations will reach out to DLP and be like, "We could use your help with this," or, "We saw you requested this. Would you request that again?" And that's super helpful. But in the future, I would love a process where lots of volunteers have ideas of different data that should be made public, and DLP is a place to find the resources, the community, the camaraderie, the help in cleaning the data to be able to do that. So that it's not just me deciding what gets requested.

This is kind of what we have right now, where if something is interesting, we have a research document. I'll share the research document with a volunteer who posts in the Slack and is like, "Hey, I'm kind of interested in requesting this." And I'll be like, "Hey, can you fill this out? And then let's talk about it in our next meeting," Or I'll talk about it one-on-one with them to make sure we have the request scoped out, and really know what we're asking for, and it's gonna be worth the time to request this thing.

Okay, so you're training a team that can then help those individual requesters on their own.

Yeah, exactly. Training a group of people who can (mostly through the Slack channel right now) collaborate on sharpening the requests, filing the requests, and cleaning the data to make it public.

That's really cool. I didn't know it was a community effort.

That's kind of what we're trying to do with the Prairie Fire Project—connect researchers with each other, and connect them with tools [and other resources]. Connecting them with this data liberation community would also be really great.

Yeah. If you have a great idea for data that you know the federal government has, or local governments have, and you're burdened by publishing papers, grading students' work, doing all the things that you do in an academic context, and you don't wanna fiddle around with FOIA and cleaning data—that's what we wanna do. That's what this group is for. The [DLP] is not just an institution that you interact with one way, it is a volunteer effort, a community.

I love third spaces. I want DLP to be a third space for people, where they can feel connected to other people who have the same ideas—in terms of open-source philosophy and culture, and civic data, and those sorts of things. So the goal is to get it to that place.

There are some really good models, like Bellingcat, for example. I'm on their Discord, and they have a good amount of volunteers who contribute OSINT research to investigations, and get bylines, and are just people across the world who are excited about this, and really find a lot of meaning in it.

We want to find those sorts of people in the public records world, and give them tools and community—things to make [this work] not disillusioning and lonely. We wanna make the process fun and useful for the people that wanna do it.

The disillusionment and the isolation are such huge problems. And collaboration is a solution to them, I think. So community building, third space building is foundational. But maybe people haven't really thought about it that way before.

Yeah, exactly. You're speaking to the thing we talk about so often at MuckRock, and that is on my mind all the time nowadays.

I think now, more than ever, there’s a need to feel that your work is connected to real things out in the world, and that it's not just falling into an internet sinkhole of AI slop. It's a real thing, it's doing real things for real people, and the internet can still be a place where that can happen.

That's beautiful. So, where can people get involved and submit their requests? You said Slack?

Yeah. Right now, the main space for volunteers wanting to get plugged in is Slack. If you go to muckrock.com/slack, you can join the public Data Liberation channel. You'll see me talking with other volunteers there, and you can just say hi, and I'll jump in and start talking to you.

I think that's the easiest way for now. We don't have so many active volunteers on any given day that I can't answer a message from someone who's like, “How do I get involved?'" So I would just answer you, or anyone directly. And if you get in that Slack, you'll see I post every week, like, "Here's the different ongoing volunteer efforts. Here's where we are. Here's where we could use help." So someone could just plug in through there.

Do you guys work with the Data Rescue Project at all?

I know of them, and actually a few of those folks are in the DLP Slack channel. And I've joined different chats or communications channels for the work that they're doing, and attended webinars. So there's some amount of overlap in communication.

When a question about data rescue comes up, I'm like, "Let's ask the experts." And then I find a way to get to them about a question, or just tag one of their directors.

I was wondering how much crossover there was between your guys' volunteers, because it seems like there would be [a bit].

Yeah, definitely. And I don't think it's something that we've fully tapped into, as much as we could. I think they've done an awesome job of just moving as quick as they could on things.

But I think for DLP, we kind of already had some ideas of what types of requests we wanted to file, what we were working on, trying to get out the door. And then 2025 started and all of a sudden we were in a completely different place. Obviously, we couldn't have known [what was coming] but we maybe didn't move as quickly to join in those efforts. We could have redirected our volunteers into some of their stuff as well.

It's really hard to coordinate. And all these datasets went down so fast. But I think merging these efforts and getting more cohesion across the data rescue space is really important. That's one of my goals.

Yeah, we need to get in front of these people and be like, "Here's what we could do. If you wanna join the project, we can help do these sorts of things. Here's the resources and skills we have, and how we can collaborate.”

I think they would all be excited to hear that. I mean, I see that all the time on the Federal Data Forum of people being like, "Hey, I built this data tool," or like, "Does anybody know how to use this [resource]?" So yeah, I think people would be really happy to hear that you guys are looking for collaborators.

When you guys uncover an interesting dataset, a really foundational one, do you have channels that you use to point reporters to it? How do you get the word out?

That's something we're still trying to figure out. I think there are a couple ways of doing it, depending on how useful and how accessible the data already is. I think there's varying degrees of data releases.

One could be we just get it up on the Data Liberation Project site with documentation. Everything we do has a pretty hefty amount of documentation. Some volunteers contribute to documenting what we know about the data, what looks fishy, what looks solid, what sort of variables are included, if there are redactions, that kind of stuff. That's the most basic—just get it on the website.

The other way would be also publishing on MuckRock at the same time and being like, "Hey, we published this data on the DLP site. Here's different ways you can use it." And maybe I, or some volunteers, speak with experts in the subject matter or field of the data, and we have a little bit of a guide of exploring it as well.

And then the very last part is [a guide] plus a tool like the Risk Management Explorer, where you can actually do something you couldn't do with this data before. But that requires a lot of time, and maybe you make something that's not actually super useful to people, or is not used. So I think the question is always, “What is gonna be useful, and what is going to be used by the most amount of people?”

So [it’s about] figuring out what level of work we wanna put into making the data accessible in some way, and then publishing it—getting it out in the MuckRock newsletter, and to our Slack, and to collaborators the best that we can.

“Transparency is a team sport… [Academic research is] meant to be as comprehensive as possible, and a lot of journalism is meant to be illustrative.”

I was looking into your article on DataJournalism.com, which is about the “Uncounted” investigation. That was part of Documenting COVID-19?

Yeah. “Uncounted” was a part of Documenting COVID-19. Documenting COVID-19 was a project that aimed to collect documents about the pandemic, as it was happening. The goal was to produce our own investigations, and also make documents public that would be useful to other newsrooms across the country.

This article on DataJournalism.com is the behind-the-scenes story—how you collaborated with a research team at Boston University, and the way that those two [approaches to research], the academic side and the journalism side, made a report possible that wouldn't have been otherwise. Is that how you would describe it?

Yeah, I think that's exactly it. I think at MuckRock, as an organization, we believe in collaboration above almost anything else. And I definitely believe that personally.

When I was in New Mexico, I was working on a lot of stories alone. Just getting as far as I could, bringing it to my editor, going through a draft, doing that traditional process. Investigative journalism takes a long time, and it's kind of lonely work. So part of the reason I was really excited for the MuckRock job, was it was essentially about collaboration. The team-sport aspect of the type of work that I'm able to do, and that we really try to do, is so much fun to me. And I think it’s so important.

MuckRock's tagline right now is "Transparency is a team sport." And, I think that's really true.

The work that we did on “Uncounted” really wouldn't have been possible without the type of research that the team of demographers at BU were doing. And they wouldn't have been able to take the big picture modeling and trends that they were seeing, and actually understand what's happening on the ground—you know, “What is the source of these numbers we're seeing? And what kind of human decisions are being made that are causing these numbers?”—without reporters who would go ask coroners how they were certifying death certificates.

So it's really about passing the baton, I think, in terms of transparency or public accountability. And that's something we really believe in, and I personally believe in.

I like that tagline a lot. So you were filling in the why for the researchers?

Yeah, the why and the how, I think. The why may have been why the coroners were deciding not to certify things as COVID that showed all of the signs and symptoms of COVID, and the person had a positive test for COVID. And the how is the process behind that, because I think that part was important too. Someone was able to just decide they weren't going to label anything as a COVID death, you know? We had a coroner who told us [he wasn’t doing] that, and he was later charged for falsifying death certificates. The how part was: This is possible in the kind of patchwork death investigation system that we have in the United States.

The cool part is the nitty-gritty details that came through in this modeling that the academic team was doing. They were looking at the difference between normal deaths, the anticipated amount of deaths for certain categories in a given year, and then the amount that were actually happening, versus the amount of COVID deaths that were happening. And they were seeing these signs that there's a gap between [anticipated non-COVID deaths and reported non-COVID deaths], and the gap is likely from things that were COVID, but were not certified as COVID. Because they were seeing a lot more deaths of different cardiac issues, a really odd amount of deaths from different cardiac issues, in a place where there were intense waves of the pandemic. So, during those time periods, is it possible that all these people were just dying of different cardiac arrests more than ever? Or are those COVID deaths that weren't being certified as such because the death investigator decided they weren't going to write “COVID” on the death certificate?

Yeah, why were they doing that? Why would you wanna put a cardiac arrest instead of COVID?

That's a good question. The thing that I heard more often than not, at least from a couple death investigators, was the family told them not to. Or their own personal belief was that people weren't dying from COVID, and they decided not to write that on the death certificate.

As a political kind of thing?

Yeah, I think.

You talked about how more newsrooms and investigators became involved that were connected to local scenes. Were these reasons, the hows and the whys, changing based on the community that was reporting [the deaths]?

Yeah, definitely.

Like, Mississippi was very different from Missouri—the different waves of the pandemic and how they played out there were different. So I think it was really essential to have reporters on the ground, who knew what it was like to live through that wave of the pandemic, and who the coroner was, how the local political systems worked, who have a better understanding of the context of the numbers we were seeing.

Wow. So you've got this national-level statistical model from the university—they're pointing you guys at the story, and then it breaks down further and further [into local teams]. I think all of that working together is really interesting.

You also talked about how, normally, an academic-journalistic collaboration is pretty one-sided. We interview the experts and they get publicity, but that's pretty much all they get from us. But the way you guys were contributing to the knowledge being generated, I think is really cool. It’s a good example of how research that's [lost] government funding can look for support from journalists, from other investigators, and from local scenes.

Yeah, absolutely. I can remember lots of conversations we had with the academic team at BU about—they wanted to know what was going on [at the local level] too. It would give their research more relevance, and more practical significance.

And I think you're right. For academic researchers, journalists can often be a way of passing the baton into a different perspective, and a different way of approaching an issue. I used [these terms] over and over again with the team: a lot of their work was meant to be as comprehensive as possible, and a lot of journalism is meant to be illustrative. You can find a balance in those things, you know?

As journalists, we're like, "We don't really need a more complex model. We need a more approachable illustration of this thing." We're trying to illustrate what this looks like from a human perspective, and maybe not rely as much on the modeling as an academic team that is trying to be as comprehensive as possible. And then you find the balance between the two things.

As we're seeing this anti-intellectualist tilt—attacks against scientists and researchers—I think being able to be understood by the public, and relate to them on a more storyteller, more emotional, and localized level is really important and, has maybe been neglected. So that also is a really promising aspect of that kind of collaboration.

Totally.

And I think we both, academics and journalists, want the same things in a lot of ways. Like one, we both want to do really rigorous, accurate, fair work. But we also have a deep and well-founded fear that even if we do really accurate, rigorous work, it might not be relevant, you know? Like, we can't spend a year on a giant investigation that no one cares about, and doesn't change anything in the real world.

“…lots of our requesters are interested in things that are local to them. So that’s the way that I think about public records...”

So you said you're kind of leaning away from FOIAs [for the Data Liberation Project]. I know your last “liberated” dataset was in fall of 2025, and I think the last request you put up on the site was [filed in] April, 2025. Do you have any active requests going on right now, or are you leaning more on web scraping? What's going on?

Yeah, that's a good question. So when we inherited the project, Jeremy had a lot of requests going. And during the first year of keeping the project alive and figuring out the direction we wanted to go, we were getting requests every couple months back from all these requests that Jeremy had filed. And so the first year was really just trying to keep up with those.

Now, I think we have filed requests at the state level in a few different cases. We haven't gotten those on the site yet because we're still getting different states back. And we'll still be filing FOIA requests at the federal level as well. But I think doing a lot more of this, [filing requests] state by state, is something that we're trying to do while still getting requests back from all those from the first couple years Jeremy and others had been filing.

So we're filing more requests at a state level. And there are more requests I would like to file, especially in the area of environmental data from the federal government. But having more volunteers who are interested in that would really help us be able to do that.

One of the bigger focuses right now is patient safety. And so we've been gearing up in the last few months to file different styles of requests to different states across the country for data sets related to patient safety, and start digging into some stuff there.

Okay, interesting. Why the switch to states? Is it a little bit easier to deal with them or...?

Yeah, I think especially right now. And I think it helps to not put all of our eggs in one basket. Like, if you have a really good, really well-thought-out FOIA request to a federal agency, that could still take a long time, especially now.

At the state level, public records laws and the compliance with them is playing out differently than at the federal level. So it's a little bit easier to get those requests back. And instead of just filing one request, we can have an idea for a request and then file it to 50 different states and then have maybe 20 of those come back. And that's 20 data sets that we can publish, rather than just one.

So we still wanna do the federal stuff, but I think also the local part is just something that MuckRock has been doing for a longer time as well, because lots of our requesters are interested in things that are local to them. So that’s the way that I think about public records, and what I’m interested in, and it helps us produce things that are useful to local reporters.

Gotcha. In the coverage of the research apparatus being dismantled, there's a lot of panic (with good reason) that these datasets are gone. But I think people don't realize that the state government also collects similar data. Your local city will sometimes collect that data.

I know Jeremy Singer-Vine, in one of the founding Data Liberation Project documents, talked about trying to craft FOIAs that are as persuasive as possible, and address typical pushback that you would get. Can you talk about what kind of pushback that would be? How do you make it persuasive and, again, has that changed since Trump?

Yeah. I think making it as persuasive as possible usually means knowing as much as you can about what you're asking for beforehand, and being able to say where exactly that data is located.

I often think of it physically, even though the data is not in a hard drive or in a filing cabinet. Like, whose filing cabinet is this? Who's in charge of this data? Where is it located? How can I describe it? If we search the agency's name and then file type— “PDF,” or “file type: PowerPoint,”—in Google, can we find training materials where the government agency is describing the dataset themselves? We can use that to describe to the FOIA officer what we're requesting. Most of the time, [making it persuasive] just means doing the research that we need to know how to describe the data, like asking for a data dictionary and other things that'll help us know more about the data. And trying to make that the least burdensome on the FOIA officer as possible.

And the type of pushbacks are usually around that—the dataset wasn't described correctly, or that this thing isn't stored in the way that we described, or that what we're requesting is overly burdensome because of X, Y, or Z reason.

You would think that, in an ideal situation, the FOIA officers would be able to parse a non-expert's request. As it's supposed to be public data that serves the layman as well as the expert. But I think staff shortages have probably made that impossible. So that makes sense as a strategy.

And then, what kind of web scraping are you guys doing?

So most of the web scraping happened prior to MuckRock taking over. There are a few that are still running on GitHub actions, I think four of them in total. One is related to firearm background checks. I can't remember offhand what the other three are. But they're kind of just running right now.

The scraping is actually not running into a lot of problems with the website changing. So we haven't actually had to update those much. And because I haven't worked with volunteers, on those and the volunteers that were working on those are prior to me [becoming director], I don't think a lot about them, I just know that they're happening.

I think some people are still using the data that they produce, but I haven't been involved in those as much.

They're just chugging away in the background?

Yeah, just chugging away in the background.

I mean, we wanna do more scraping with volunteers. But at this point, I think it'd be up to a volunteer to bring that idea forward and be like, "I think we should scrape this, or do that."

On the other hand, the internet is always changing, so I think going after public records requests that are a single stagnant dataset might be a better use of our time. Not to say that scraping is not important. It definitely is, and will always have a place. But they've just become projects that present a lot of different challenges, especially for volunteers that are doing this on their nights and weekends outside of their regular job.

Gotcha. You don't wanna point them to a website that's not there anymore.

Yeah. Or it could be six months and they restructure the whole website and then the scraper falls apart or the parsing part falls apart.

“…the goal is that it all feeds back, both to our work and to other collaborations.”

Have you, or has MuckRock, ever done an investigation based on the Data Liberation Project data?

No, not yet—not built a whole investigation around the data. We've seen other newsrooms use the data, or data being cited in different places. But the goal is to help local newsrooms.

And we have a research helpdesk. So if we're doing state-by-state requests for data that volunteers and newsrooms help liberate, then the research desk can get that to them and be like, "Hey, we just also cleaned up this data about adverse health events in hospitals in your state. We know how the data works. If you want us to do a training on it, we can do a training for you as well." So, the goal is that it all feeds back, both to our work and to other collaborations.

Okay, so you already have a journalistic outlet, an on-the-ground outlet in mind for a lot of this.

Yeah. I think we have enough Sunlight Research Desk member newsrooms that we can reach out to, directly.

Is there anything I missed?

One thing I would love to highlight is one of the cool projects I'm really excited about.

The Marshall Project had been requesting banned books lists from state prison systems across the country for a while, and in the last year or so we were talking with the Marshall Project—just kind of in general trying to figure out a way to collaborate because we'd been friends as organizations for a while—and they're like, "You know, one thing we could really use help with is refreshing the data for banned books." And we were like, "That's great.” This is a great style of project for DLP volunteers, because this is something very practical, and I think really immediately, understandably important to volunteers, even if they're not so familiar with the criminal justice system.

So in the last few months, I filed some more requests and now I'm working with volunteers on cleaning those datasets. And then, leading up to Banned Books Month in October, the volunteers are gonna be helping take what the Marshall Project did a few years ago, the database that they made, and refresh the data. So I filed the request, and now we have a GitHub where each state is an issue, and then volunteers are tackling each state.

Hopefully it'll lead to the Marshall Project being able to get new data out to reporters in those states. If you're in Texas, or California, or any of the other states that we've been able to get this data on, you now have a more recent list of books that have been banned in prisons. And you could go talk to people in the prison system about this, or talk with authors whose books have been censored.

We actually haven't figured out the logistics yet, but I think they'll still host the data, maybe not in the same way they were before. We're still trying to figure out how we also take these clean lists of books and highlight the ones that are the most important for people to know have been censored, because they're often big lists of different types of books.

And then they also have this project called Investigate This!. The idea is to get datasets to reporters, and guides to reporters on how to use that data. So hopefully this will work its way back into an Investigate This! guide.

Right, like what you were saying with the COVID project—it's the groundwork for the storytellers and the local investigators to then pick up. So that's awesome. I'll be looking out for that.

Okay. Awesome. I think I got through everything. Thank you so much for your time.

Yeah, it's been a great conversation. I'll see you in the DLP Slack.