The honest answer is sometimes, and less often than most teams assume.
The question usually gets framed as a capability problem. Can our DevRel team make good videos? Almost always yes. Developer advocates are frequently excellent on camera, they know the product better than any agency will, and they can answer a question mid-recording that an outsider would have to go and research.
That's the wrong question. Paul Bratslavsky, Developer Advocate at Strapi, put the right one better than I can:
At Strapi, the challenge was never that we couldn't make videos ourselves. It's that we probably shouldn't. Video is a whole pipeline. You have to source topics, write scripts, record, edit both long-form and short-form, and then publish. That's a full-time craft, and it deserves people who do it full-time.
the real cost is the hour, not the video
A developer advocate's most valuable hours go into deep technical content and time with the community. Those are the hours nobody else in the company can spend. Editing a twenty-minute build video into shorts is not one of those hours, and it eats four of them.
So the calculation isn't "can they" or even "what would an agency cost." It's what the DevRel team stops doing while they edit.
when in-house is right
Keep it in-house when the videos are a small part of a bigger job that only your team can do. Conference talk recordings, live streams, release walkthroughs where the person on camera needs to answer a live chat question about the roadmap. Anything where the value is the specific person, not the production.
Keep it in-house if you have someone whose job title actually is video. If video is somebody's whole job rather than the fifth thing on their list, you don't have the problem this post is about.
when it isn't
Hand it off when you need cadence. A pipeline that has to produce something every week regardless of what else is happening is exactly the thing that breaks when it's owned by somebody with five other responsibilities.
We looked at 100 funded devtool companies in July and 48 had no findable channel at all. Very few of those companies decided against video. They just never had anyone with the space to keep it running.
the middle option nobody names
You don't have to choose the whole thing. The part that reliably kills momentum is post-production, not filming. Plenty of teams keep their advocate on camera, because that's where the credibility is, and move sourcing, scripting, editing, and publishing somewhere else.
That way the face of the channel stays yours, and the four hours of editing stop landing on the person you hired to talk to developers.