What this guide helps you do
Create realistic historical reconstruction scenes with AI
Key takeaways
- Research the physical world before writing cinematic prompts.
- Use an image-first anchor and simple motion for each shot.
- Frame realistic results as reconstruction, never archival evidence.
Film evidence
Reconstruction should make its evidence boundary visible
The film stages a living Pompeii and the eruption through generated scenes. Those images can orient a viewer and support a sourced narrative, but they combine documented event, historical inference, production design, and model invention. The edit must not collapse those categories into apparent eyewitness evidence.
Write a visual evidence brief before generation: identify what each scene needs to communicate, which details are supported, which are bounded inference, and what must remain generic or off-screen. After generation, review every realistic specificity and add disclosure near the film. Cinematic plausibility is an editorial tool, not provenance.
Pompeii: Vesuvius Erupts · 01:01 · Historical reconstruction. This is a finished first-party Onira production, not customer proof or archive footage.
View the full film and production notesWhat to watch for
- Source-backed historical anchors in the narration
- Invented specifics that should be removed or qualified
- Explicit identification of generated reconstruction
Section 1
Build a material-culture reference sheet
Before generation, collect reliable references for architecture, street layout, clothing, tools, furniture, transport, vegetation, weather, light sources, social roles, and signs of wealth or labor. Record which details are documented and which are informed inference.
This sheet should be specific to a place and period. Generic medieval, ancient, or wartime styling produces familiar imagery but weak historical communication.
- +Location and date range.
- +Materials, construction, clothing, and tools.
- +Environment, season, light, and social context.
Section 2
Direct a still image before movement
Describe subject, setting, spatial relationship, camera position, lens behavior, light, color, texture, and historical constraints in the still prompt. Generate and review the anchor frame before spending motion credits.
A strong anchor exposes anachronism and composition problems while they are cheaper to correct. It can also become a reference for connected shots or recurring people and places.
- +One clear focal subject.
- +Period and geography constraints stated explicitly.
- +No readable generated text used as evidence.
Section 3
Keep motion physically simple
Ask for one main action and restrained environmental movement: a craftsperson turns a tool, dust moves through window light, a cart passes at a distance, or the camera slowly reveals a street. Complex crowds, handoffs, choreography, and multiple interacting bodies increase failure risk.
The motion should serve the narration rather than demonstrate everything at once. A sequence of specific short shots usually communicates more clearly than one overloaded generation.
- +One principal action.
- +One camera behavior.
- +Subtle environmental motion.
Section 4
Review historical and documentary truth
Inspect the scene with historical references open. Look for anachronistic materials, impossible geography, modern facial styling, incorrect insignia, invented writing, wrong tools, or social behavior unsupported by the evidence.
Then ask how the viewer will understand the image. Use disclosure and contextual language where needed, and avoid visual treatments that imitate authentic archive or eyewitness footage. Reconstruction should clarify the story without overstating certainty.
- +Historical specialist or source-based visual review.
- +No fabricated evidence or false archival framing.
- +Disclosure decision recorded for realistic scenes.
Working standard
Publication checklist
- 01A period-specific material-culture sheet exists.
- 02Documented detail and informed inference are distinguished.
- 03The anchor still passes historical and compositional review.
- 04Motion uses one principal action and restrained camera direction.
- 05Generated text, maps, insignia, and likenesses are checked.
- 06The scene is framed transparently as reconstruction.
Primary references
Sources and further reading
Policy and model capabilities change. These sources were reviewed on July 11, 2026; open the current official page before making a production or publication decision.
Related production guides
Questions
What is AI historical reconstruction?
It uses generative image or video systems to illustrate a past setting, person, object, or event. It remains a reconstruction and should not be presented as authentic footage or documentary evidence.
How do I reduce anachronisms?
Research material culture, constrain period and geography in the prompt, review a still before motion, and compare the result with authoritative references before publication.
Can reconstruction replace archival material?
It can illustrate scenes for which footage does not exist, but it cannot provide the evidentiary value of an authentic document, photograph, recording, artifact, or eyewitness source.