How do you generate a reverse angle shot with AI video when you have no reference image?

You generate a reverse angle shot with no reference image by keeping the scene's spatial geography in an AI agent's session context — the invideo agent reconstructs the opposite side from prior shots instead of an image input. Four ways to do it:

1. Reverse from established scene geography
2. Ask the invideo agent to surface the unbuilt side first
3. Chain the opposite angle off your hero shot
4. Lock one world element and extract every angle

Work the reverse angle inside one continuous session, because the spatial information that replaces your missing reference image lives in the conversation: prior shots, character positions, and the axis of action your coverage has to respect. invideo is an agentic video creation tool with all the current models (Veo, Kling, Seedance 2.0) available, and the invideo agent holds that scene geography across the session — which is what makes reference-free reverse angles possible, where most camera-angle workflows require a source clip or master shot to re-angle.

1. Reverse from established scene geography. Generate your forward coverage first, then ask the invideo agent directly for the reverse on your character. Because it retains the geography established in earlier shots — room layout, character placement, lens grammar — it can build the opposite side with zero image input; in one documented production the invideo agent reconstructed a precise reverse angle from session geography alone, holding character, lighting, and spatial logic across the cut. With that context loaded, a minimal continuation prompt like "Everything should match" is enough to keep the reverse consistent with the shot it cuts against.

2. Ask the invideo agent to surface the unbuilt side before generating. A reverse angle points the camera at a part of the set your forward shots never showed, so instruct the invideo agent to apply art-director logic rather than simple mirroring: have it flag undecided production design elements and propose narrative-loaded options. In one session the invideo agent responded to a reverse request with "Reverse on Marcus — what's behind him? That near wall doesn't exist yet. What should it be?" — you pick the option, then it builds the frame, which stops the model hallucinating an off-axis background.

3. Chain the opposite angle off your hero shot. Once you have a generated shot you like — itself made from prompts, no reference needed — immediately request the compositionally opposite angle in the same conversational session to build a matched coverage pair for editing. In one documented production this landed a complex angle on the first generation attempt after manual prompting had failed to produce it.

4. Lock one world element and extract every angle. Lock a single element of the scene with the invideo agent and it autonomously generates wide, close, and side coverage without you requesting each angle individually; pull your reverse from that set, and use the extracted frames as continuity anchors for every subsequent shot in the scene.

Whichever method you use, keep character sheets loaded in the invideo agent's context so identity holds across the cut — one 70-second production kept two characters consistent across every scene this way, no LoRA — and budget iterations, since documented productions averaged 3 generations per usable shot.

These are some of the ways to problem-solve this — what works depends on your scene.

Watch some of these to see what works for you:

it delivered a precise reverse angle without a single reference from me. That's pretty insane, right?

— invideo's creative team, on agent spatial reasoning across a multi-shot production