How De Correspondent brought video onto its own platform

De Correspondent is an independent, member-funded journalism platform based in the Netherlands. Its correspondents publish in-depth journalism that looks beyond the daily news cycle and helps readers understand the systems shaping current events. The platform is entirely ad-free, a principle that does not sit comfortably with relying on advertising-driven technology platforms for video.

Recently, De Correspondent began publishing video directly on its own website and app. Members can now watch its journalism without ads or tracking cookies, and without needing an account on a third-party social platform.

In its own announcement about moving video away from Big Tech, De Correspondent explains the editorial and political reasoning behind that decision. It also introduced the result in a fitting way: with a short video published directly on its platform.

I spoke with Vincent, a full-stack developer at De Correspondent, about the implementation, the decisions his team made and the trade-offs they encountered.

This interview has been edited and condensed for clarity.

Could you introduce yourself and your role at De Correspondent?

I’m Vincent, and I’ve been working as a developer for around fifteen years.

I started without a formal web development education and worked at a design agency for a long time, mostly on campaigns and projects for organisations working in areas such as nature and conservation.

Later, I studied at Codam in Amsterdam, where the focus was much more low-level. I worked with C, algorithms and even Assembly. That was very different from web development, but I learned a lot from it.

I joined De Correspondent around two and a half years ago as a full-stack developer. I wanted to work somewhere with a mission I could believe in, and I’m very happy with the team we have built here.

How has the role of video changed since you joined?

When I started, video was used occasionally, but there wasn’t yet a dedicated video team or a clear strategy for how video should be used on the platform.

Funding from a foundation gave De Correspondent room to experiment with video as a journalistic format. Since then, we have created a video editorial team, acquired equipment, built a recording space and developed a broader video strategy.

One of the formats we recently launched is short, vertical video. It resembles what people are familiar with from TikTok or Instagram, but most of our videos are connected to a full publication. The short video can introduce an idea, while the article provides the depth.

Why was it important to publish those videos on your own platform?

There were two sides to that.

First, our user research showed that there is a real demand for video. That demand wasn’t limited to people in their twenties. At the same time, some members told us that they find interfaces with lots of movement distracting. That encouraged us to think carefully about autoplay, motion and the amount of control we give users.

Second, people told us they liked watching our videos on social media, but found it difficult to find them again or share them later. Publishing the videos on our own platform gives them a permanent place alongside the rest of our journalism.

Social media can still be useful for distribution, but it shouldn’t be the only place where members can access the work they are already paying for.

How did you approach the implementation?

We started with a prototype and integrated Mave relatively early.

We didn’t just want to paste a player embed into a webpage. Video had to fit into the complete publishing workflow, from editing a publication in our CMS to displaying it correctly on the website and in the app.

The initial integration was quite straightforward because Mave returned the information in a format our existing integrations already understood. We could add support without inventing an entirely new workflow.

In our editor, journalists could already paste a YouTube or Vimeo link into a video content block. We extended that block so it also recognises a Mave embed or video ID and retrieves the relevant metadata.

For now, uploads still happen inside Mave. Editors then use the video ID in our CMS. That was close to the workflow they already knew and kept the first implementation manageable.

How much did you use out of the box, and how much did you customise?

For the first beta, we used Mave Clips for the preview tiles and the standard Mave Player for playback. That allowed us to get a working version in front of members quickly.

For regular videos inside publications, we still use the standard player with very little customisation.

The vertical video feed was different. We wanted an experience that felt more like a native short-form feed, including our own controls and sharing interface. For that part, we removed the standard player interface and built a custom layer around the video.

That flexibility was important. We could start with existing components, validate the idea and then customise the parts where the product required something more specific.

Did you test the experience with members?

Yes. De Correspondent has a member panel of around 2,000 people who have said they are willing to participate in research.

We invited members to test the prototype over video calls while sharing their screens. More than 150 people volunteered for the test period, so we had a very engaged group to work with.

Those sessions exposed assumptions we didn’t realise we were making. Sometimes people expected an element to be clickable when it wasn’t, or navigated differently from what we had anticipated.

We also heard from members who said this was exactly what they had been looking for: they enjoyed the videos on social media but wanted a place where they could reliably find and share them.

Was there a less visible technical problem Mave solved?

Moving images were a surprisingly good example.

Historically, those were GIFs. Later they became MP4 files, but each one was still a single file with one resolution and codec. Sometimes a 4K video would end up being downloaded on every phone, even when it was displayed at a much smaller size.

There were also device-specific problems. A video slightly above a supported resolution could simply appear as a grey area on some older phones. When that happened, the issue quickly reached the development team, and we ended up debugging individual MP4 files.

Now the editorial team can upload the best source file they have, and Mave creates the required versions. They no longer have to think about resolution, codec or bitrate for every use case. That removes a category of support work from the development team and gives editors more freedom.

Digital sovereignty was important, but were there other reasons for choosing Mave?

It was an important factor, especially because De Correspondent is working more broadly to reduce its dependence on large American technology companies.

But ease of development mattered too. I was enthusiastic about the components because I could request a clip or a player and much of the required behaviour was already there. We still built a substantial amount ourselves, but we would have had to build even more on top of a lower-level video provider.

We also wanted to launch a beta quickly. We needed a good player, and Mave already had one.

That doesn’t mean there were no concerns. Mave is a relatively small company, and its pricing model isn’t necessarily ideal for every part of a permanent journalistic archive. We discussed those risks and accepted them.

One thing that helped was knowing that the videos remain ours and that we can access the original files. If a provider disappears or changes direction, we don’t want our entire archive to disappear with it.

Does this mean De Correspondent no longer needs YouTube or social media?

Not entirely.

YouTube and other platforms are still useful social channels. They provide discovery, comments and access to audiences who don’t know De Correspondent yet.

The distinction is that we no longer want the YouTube embed to be the default player on our own website. On our platform, it adds little for members while introducing an external platform into their viewing experience.

That creates a trade-off: views on our own player do not contribute to YouTube’s recommendation algorithm. But it also means our journalism remains available regardless of what an algorithm decides to promote, or what a platform decides to remove.

What has the response been?

We have received many positive responses through different channels. Some members contacted our member team without being prompted, simply to thank us.

It’s also helpful that video was a new area for us. Instead of migrating a deeply embedded existing system, we could build in the direction we already wanted to go.

There is still more to develop, but we have already taken a significant step.

Digital sovereignty is an engineering decision

What I like about De Correspondent’s implementation is that it doesn’t present digital sovereignty as an absolute or a slogan.

The team still publishes on social media. It made pragmatic choices about its CMS workflow. It accepted some risks and rejected others. It used standard components where possible and built custom interfaces where the product needed them.

That is what technological independence usually looks like in practice: not disconnecting from everything overnight, but deciding which parts of your infrastructure need to remain under your control and then building accordingly.