Recording platform Riverside, used by Spotify, NBC and the New York Times among others, introduced Riverside Agents in late October 2026: a team of artificial intelligence agents designed for video production. The tool is available now on web and mobile, on every subscription plan, although some capabilities depend on the plan tier.
Four roles to stand in for a production team
Riverside's starting point is simple: video has become the default format for podcasters and creators, and small teams have to handle preparation, editing, clipping and distribution on their own. The company has therefore split the production chain between four specialist agents:
- The Producer prepares the session: interview questions, talking points, scenes, layouts and branding.
- The Editor turns the full recording into a finished episode. It suggests where to tighten the conversation and which moment should be the hook, then adds layouts, B-roll, motion graphics, captions and audio polish.
- The Clipper spots the strongest moments and turns them into clips ready to post.
- The Marketer writes the accompanying copy, creates thumbnails, decides where each clip should run and schedules the posts.
Agents that move on their own, but under control
The difference Riverside claims from conventional AI tools is autonomy. Until now, a creator wrote a prompt, checked the result and moved the work to the next step themselves. With these agents, the work carries on once recording ends: when the host checks back in, an edit, clips and posts are waiting for approval. Riverside stresses one point: nothing goes live without the creator's consent, and the creator decides what to hand off. Over time, the agents are meant to learn the creator's preferences and editing style.
Each agent relies on a proprietary model trained for one specific job, along with Riverside's in-house tools for editing conversations, such as removing pauses and filler words. According to the company, this specialisation uses less computing power and lets creators finish full episodes without hitting a usage limit. Everything runs in the cloud, so you can close your laptop and approve clips from your phone.
Founded in 2019 by brothers Nadav and Gideon Keyson and based in Tel Aviv, the company says more than 40 million sessions have been recorded on its platform. "Every creator deserves a production team," sums up its CEO, Nadav Keyson.
What it means for webradio stations
For a webradio run by a handful of volunteers or staff, the main obstacle to video is not the camera but the time spent in post-production. Filming a show in the studio has become easy; editing it, cutting short formats for social media and publishing them every day is much harder. Tools like this could help a small station turn a breakfast show or an interview into a video episode and a set of clips without growing its team.
A few points still call for caution:
- Keep editorial control: choosing the hook or the passages to cut reflects the station's identity and should remain a human decision.
- Check the plan: some features depend on the subscription, which matters for a small organisation's budget.
- Protect your sound and visual identity: automated branding must stay consistent with the station's on-air imaging.
The arrival of agents able to chain tasks together confirms a trend: audio and video production is moving towards a largely automated pipeline in which the host becomes the show's director. For webradios, the challenge will be to use the time saved for content, without losing what makes them distinctive.
Source: Podnews