Should you run a storyboard agent before a cinematography agent in your AI film pipeline?

Yes — run the storyboard agent before the DOP agent in almost every case. The storyboard agent locks framing, blocking, and shot order so the DOP agent inherits a visual contract instead of inferring camera, lens, and lighting from text alone. The only exception: very short or exploratory pieces, where you can skip ahead and treat early DOP output as a rough storyboard.

Set up the pipeline in this order: creative producer agent → storyboard agent → DOP agent (and the rest of the crew downstream). invideo is an agentic video tool where you spin up named sub-agents inside one project, each holding scoped context, so this sequence is something you actually construct — not a built-in button.

1. Initialize a creative producer agent first. This agent holds the full script, shot breakdown, character details, and your visual language reference for the whole film. Every downstream agent is grounded against it, which is how you avoid each specialist agent re-interpreting the project from scratch. As Hridaye, invideo's creative director, puts it: "To really set up the context for the agent, I normally start off with the creative producer agent. That's where I'll give the script, or the shot breakdown, along with the characters. That's the main agent that sort of holds the understanding and the vision of the entire film."

2. Run the storyboard agent next. Before any cinematography direction, have a dedicated storyboard sub-agent visualize each shot — framing, blocking, shot order, beat-to-beat coverage. This produces the visual brief the DOP agent will work against. Skipping this step forces the DOP agent to guess at composition and continuity from prose, which is where style drift and broken coverage start. Production logic frameworks like StoryPipe's 8-stage pipeline and the Camera Artist research model both place storyboard generation upstream of cinematography for the same reason.

3. Then bring in the DOP agent — often more than one. Assign a DOP agent per scene rather than one DOP across the whole film, because each scene wants its own lens grammar, lighting plan, and movement vocabulary. On complex scenes, run two DOP agents on the same sequence in parallel to bring multiple visual perspectives. In one documented production a single creator ran 6 agents simultaneously; another ran 8 specialist agents across separate project pages — that scale is only manageable because the storyboard has already locked what each DOP is shooting.

4. Direct, don't prompt. With the storyboard locked, the DOP agent receives on-set-style language: "I want to stay on the feral guy when we run this scene. No back and forth cutting. We hold on him right up till he lunges" — and it executes against a visual brief, not a blank prompt. invideo holds all the current video models (Runway, Veo, Kling, Seedance 2.0) and routes each shot to the right one, so the DOP agent's job is composition and coverage decisions, not model selection.

The exception. For pieces under ~3 scenes, highly iterative R&D, or single-shot experiments, you can skip the storyboard agent and go straight to a DOP agent, treating its first outputs as a rough storyboard you refine. Once you cross into multi-scene work where visual consistency matters across shots, always storyboard first.

Decision rule. If style consistency across scenes matters → storyboard agent before DOP agent, always. If you're exploring a single look or beat → DOP-first is faster.

Watch some of these to see what works for you:

To really set up the context for the agent, I normally start off with the creative producer agent. That's where I'll give the script, or the shot breakdown, along with the characters. That's the main agent that sort of holds the understanding and the vision of the entire film.

— Hridaye, invideo's creative director