Four AI agents to produce a video, from recording to publication

Riverside deploys four AI agents to prepare, edit, clip, and publish videos, featuring automated coordination and human validation.

The recording ends. Before its creator has even looked at the footage, one agent has identified key highlights, another is preparing the edit, and a third is anticipating their distribution on social media. This is the workflow offered by Riverside Agents, a suite of four specialized assistants that divide up the different stages of video production.

The platform best known for recording podcasts and video interviews now centers its environment around a virtual team: Producer, Editor, Clipper, and Marketer. Each agent has its own features, can hand off its work to the others, and operates directly within Riverside's tools. The user retains final approval over all edits and publications. A production pipeline split among four agents

The structure of Riverside Agents reflects the main roles in audiovisual production, adapting them for in-app software operations.

The Producer steps in before recording begins. It can prepare an interview, suggest questions, organize a session's flow, and set up the studio based on user preferences. During recording, it can also draw on the live transcript to suggest follow-up questions or prompts.

The Editor then takes over the editing process. It analyzes the recording, proposes an edit plan, and can adjust the pacing of the conversation, remove silences and filler words, tweak camera framing, and add graphic overlays. Available features also include audio enhancement, as well as adding captions, music, and visual elements.

The Clipper focuses on short clips. Working from a full episode, it identifies segments likely to perform well on their own, then prepares videos tailored for social media. In particular, it can add captions, apply brand templates, and enrich clips with music or visual elements.

Finally, the Marketer handles distribution. It writes descriptions, suggests titles, generates thumbnails, and schedules content across connected platforms. Its scope also covers publishing podcast episodes and tracking select audience metrics.

These four roles previously existed in part as standalone recording, editing, or publishing tools. Grouping them into a team of agents primarily changes how users interact with them and coordinate their tasks. Agents pass work along without waiting for a new prompt

A key feature of this setup is the automatic triggering of certain tasks.

When a recording ends, the agents can immediately begin analyzing it and suggesting the next steps. The Clipper identifies highlights while the Editor prepares a draft edit. The selected content can then be handed off to the Marketer, who suggests where and how it should be posted.

As a result, users no longer need to launch each function separately. The agents can prepare an output or recommendation before receiving a new instruction.

Riverside also notes that the agents feature a memory for working habits. Over time, they can learn editing, design, or distribution preferences to cut down on repetitive settings adjustments.

However, this automation does not give the agents blanket permission to act. According to the official documentation, their proposals must be approved before being applied or published. Preparing the work and final execution remain two distinct steps. Conversational editing—with full manual control

The conversational interface plays a central role in how the system operates. Rather than manually navigating through menus, a creator can ask the Editor to shorten an intro, remove filler words, or improve the pacing of a scene.

The agent proposes a series of edits and can apply them once approved. It works directly inside Riverside's editing suite, keeping all changes fully adjustable within the project timeline.

The same principle applies to social media clips. The Clipper can generate Magic Clips, offer several options from an episode, and then take prompts to adjust reframing, add graphic assets, or tweak a segment.

Riverside's claimed specialization focuses heavily on conversational videos, interviews, video podcasts, and talking-head presentations. The documented features cater specifically to these formats, rather than replacing the full scope of a traditional video post-production suite.

In its announcement, the platform did not provide independent benchmarks measuring the time saved by these agents compared to manual editing or alternative automated solutions. Prepare and publish directly from an iPhone

All four agents are accessible from Riverside's web version and its iOS app.

A user can finish recording on a desktop computer, review edit suggestions on their phone, preview generated clips, and approve a post without returning to their workstation.

The Marketer handles several distribution tasks: writing platform-specific captions, scheduling posts, creating thumbnails, and managing episodes hosted on the platform.

The Marketer documentation notes that social media accounts must be connected before scheduling or publishing content. Every post or episode requires user confirmation.

This mobile availability is currently limited to iOS. Riverside makes no mention of a concurrent Android rollout for these new agents in its launch documentation. Available across all plans, with varying limits

Riverside announced that the agents are available across all subscription tiers, though their capabilities depend on the plan selected.

Certain operations remain tied to paid tiers. For instance, accessing the Producer directly inside the studio is reserved for Grow, Webinar, and Business plans. Tools for generation or enhancement that rely on AI credits will continue to consume them when triggered by an agent.

The company emphasizes the ability to seamlessly transition between editing, clip creation, and publishing without exhausting a conversational prompt quota at every step. However, this does not mean all operations are unlimited or free.

The agents can be tested on Riverside, which also offers a 14-day trial. The documentation covers the specific capabilities of the Producer, Editor, and Clipper