Riverside connects four AI production agents, with approval before edits and posts

Four production jobs share a recording workflow

Riverside is dividing the work around a video recording among four AI agents: Producer, Editor, Clipper and Marketer. Its October 27 issuer release follows an official product introduction dated October 26. For podcasters and small video teams, the useful change is the connection between preparation, editing and distribution inside the same recording platform.

The Producer prepares questions, talking points and studio elements such as scenes, layouts and branding. The Editor proposes changes to the full recording and an editing plan, while the Clipper turns selected moments into short videos. The Marketer prepares captions and thumbnails and can publish or schedule the resulting content after approval, according to the release. These are Riverside’s descriptions of the roles; OpenTools has not independently tested their editorial judgment.

A handoff does not remove the approval step

The official overview describes delegation within the conversation. If a request belongs to a different specialist, an agent can hand it over: the example is asking Producer for Magic Clips and having Clipper take the task. The point is to keep the request moving without requiring the creator to select the next agent manually.

Riverside says the agents can begin preparing work after a recording uploads. Its introduction states that they do not apply changes or post content without the creator’s approval. The release similarly describes edits, clips and posts waiting for a decision. That makes review a separate stage of the workflow, rather than treating an agent handoff as permission to release a video publicly.

For a creator, the approval boundary is where proposed cuts, captions and publishing choices still need to be checked against the recording and intended destination. Riverside’s stated approval design is not evidence that a particular generated edit or post will be accurate.

Studio context and settings carry across the agents

According to the help documentation, each agent keeps an ongoing chat and remembers the conversation and current work. Context is held at the studio level. The Editor and Clipper can use saved branding, including a logo, background and caption style.

Tone and workflow preferences apply across all the agents; the documentation says they cannot be set separately for each one. A team choosing its studio settings is therefore also choosing shared instructions for the production chain. A different caption style or workflow for one specialist cannot simply be assumed to have an independent settings scope.

Web and iOS access still have plan and credit limits

The official introduction documents access on the web and in Riverside’s iOS app. Agents can be opened from the Home page or alongside work in areas such as Planner, Editor and Hosting. On iOS, they are reached through the Agents tab. This source does not establish Android availability.

The issuer release says Agents are available across Riverside plans, while noting that some capabilities depend on the plan tier. Separately, the overview distinguishes conversation from paid work: chatting with an agent does not use AI credits, but a task that costs credits shows its price and asks before spending them. Access to the agents should therefore not be read as unlimited use of every production capability.

For someone evaluating the workflow, the practical questions are which tasks their plan includes, what a proposed task costs and what they are approving before it changes an edit or reaches a publishing destination. The four‑agent design connects those stages, while leaving those decisions with the creator.

AI‑generated conceptual illustration by OpenTools Team: a microphone, film‑strip timeline, approval gate and publication cards represent the recording‑to‑release workflow. This is illustrative artwork, not Riverside’s interface, an actual recording or evidence of a tested approval control.

Share this article

Related News

Cato AI Endpoint brings AI traffic and local-agent controls alongside existing security tools

Easy MCP AI 2.3.0 expands WordPress editing tools and mid-task permission requests

Zendesk Auto assist permits up to seven pre-approved action executions per suggestion

Riverside connects four AI production agents, with approval before edits and posts

Four production jobs share a recording workflow

A handoff does not remove the approval step

Studio context and settings carry across the agents

Web and iOS access still have plan and credit limits

Sources

Tags

Riverside connects four AI production agents, with approval before edits and posts

Four production jobs share a recording workflow

A handoff does not remove the approval step

Studio context and settings carry across the agents

Web and iOS access still have plan and credit limits

Sources

Tags