How to Archive an AI Vertical Drama Production for Sequel Use: What to Save, How to Store It, and What You Will Need in 18 Months

The sequel commission arrives 18 months after the primary series delivery. The production team opens the archive to find the character reference pack. What they find instead is a folder of images without naming conventions, a generation log that has not been version-controlled, and a style guide document that references a character model ID that no longer exists on the current generation tool's infrastructure.

The sequel begins with a complete pre-production rebuild rather than with the infrastructure the primary series built. The rebuild costs two to three weeks. The sequel is delayed. The platform that was expecting the sequel at the 12-week mark receives it at the 15-week mark.

This scenario is not hypothetical. It is the outcome of treating the production archive as a storage task rather than as a franchise infrastructure task. The production archive's commercial function is not documenting what was produced. It is preserving the capability to produce the sequel at the efficiency that the primary series' production cost assumed.

This post covers the complete archive structure for AI vertical drama production: every category of material that must be preserved, the specific file format and naming convention for each category, and the storage structure that makes the archive usable rather than merely complete.

The Archive Mindset Shift

Most production archives are backward-looking: they preserve what was produced for contractual, legal, and audit purposes. The AI vertical drama production archive must be forward-looking: it preserves what the sequel production will need, specified in the format the sequel production can use.

The backward-looking archive stores: final delivered files, contracts, correspondence, and a generation log that records what was done.

The forward-looking archive stores everything in the backward-looking archive plus: the exact configuration that produced the approved character reference images, the Soul ID model identifiers or equivalent character model parameters, the ControlNet reference library with the geometry specification that produced the approved camera positions, the style guide at the visual specification level that the generation operator used rather than the creative description level that the brief contained, and the generation log in a format that a new generation operator joining the sequel production can read and immediately understand.

The difference between a backward-looking archive and a forward-looking archive is the difference between an archive that documents the primary series and an archive that enables the sequel.

Category 1: The Character Reference Pack Archive

The character reference pack is the most commercially important archive category. It is the foundation on which every generation session in the sequel will be built. If the character reference pack is not archived correctly, the sequel production must rebuild the character visual identity from scratch — which produces visible character drift between the primary series and the sequel regardless of how carefully the rebuild is attempted.

What to archive:

The approved reference images for each primary character at each arc position. File naming convention: CHARACTER-NAME_ARC-POSITION-NUMBER_APPROVED-DATE. Example: ALPHA_ARC1_2026-04-15.jpg. The naming convention must be consistent across all characters and all arc positions.

The Soul ID model identifier or equivalent character model parameter for each character. Store this as a text file in the same folder as the character's reference images. File name: CHARACTER-NAME_MODEL-ID.txt. Contents: the tool provider, the model identifier, the training dataset description, and the date the model was trained.

The session-open consistency check reference frame for each character. This is the single approved reference image that each production session's first test output is compared against. File name: CHARACTER-NAME_CONSISTENCY-REFERENCE.jpg. This file is the most important single file in the character reference pack archive because it is the standard against which the sequel's first generation session is evaluated.

The arc position wardrobe description for each character at each arc position. Store this as a text file adjacent to the wardrobe reference images. File name: CHARACTER-NAME_ARC-POSITION-NUMBER_WARDROBE-SPEC.txt. Contents: the specific wardrobe items, the colour palette, the styling details, and the arc position transition triggers.

Where to store it:

A dedicated character archive folder in a cloud storage service with version history enabled. Google Drive or Dropbox with version history preserves previous versions of every file. If a reference image is accidentally replaced with an incorrect version, the version history allows restoration of the original. The character archive folder must be backed up to a second cloud storage service at delivery and at 90-day intervals thereafter.

Category 2: The Generation Configuration Archive

The generation configuration archive preserves the exact technical parameters that produced the approved outputs. Without this archive, the sequel production's generation operators must reconstruct the configuration through testing rather than loading it from documentation.

What to archive:

The workflow.json file from every ComfyUI session used in pre-production reference pack building. This file reproduces the exact character reference build configuration without reconstruction. File name: CHARACTER-NAME_REFERENCE-BUILD_SESSION-DATE.json. Store all workflow files in a dedicated ComfyUI archive folder.

The ControlNet geometry reference library. The OpenPose skeleton or depth map references for each standard camera position — authority close-up, proximity close-up, vulnerability close-up, wide establishing. File name: CAMERA-POSITION-TYPE_CONTROLNET-REFERENCE.png. These files are the camera geometry standard that the sequel's generation operator uses to reproduce the primary series' camera positions.

The five-layer prompt templates for each scene type. File name: SCENE-TYPE_PROMPT-TEMPLATE.txt. Contents: the complete five-layer prompt structure that was used for that scene type in the primary series, with the character-specific parameters filled in and marked with the character's name. The sequel operator who loads this template has the approved starting point for each scene type rather than rebuilding the prompt from the brief.

The generation parameter defaults for each tool in the production stack. File name: TOOL-NAME_GENERATION-PARAMETERS.txt. Contents: the model version, the sampling method, the CFG scale, the step count, and any other tool-specific parameters that were used in approved generation sessions.

Where to store it:

A dedicated generation configuration folder adjacent to the character reference pack folder. The generation configuration folder and the character reference pack folder together are the core of the production's technical archive. Both must be backed up simultaneously.

Category 3: The Creative Document Archive

The creative document archive preserves the production's creative specifications at the level of detail that the sequel needs for brief development and arc map construction.

What to archive:

The complete arc map from the primary series. Not a summary. The full arc map document including the paywall episode position, the midpoint reversal episode and its premise reframe, the escalation arc across all three act blocks, the antagonist arc with the specific injustices at each escalation marker, and the resolution sequence's specific events.

The character bibles for all primary characters, at the full specification level including the psychological configuration, the tag stack, the involuntary tell description, the visual specification at technical detail level, and the arc position wardrobe states.

The style guide at the generation specification level. Not the mood board. The colour temperature descriptions, the lighting direction specifications, the environment category descriptions, and the ControlNet camera geometry specifications that were used in generation rather than the brief's creative descriptions that inspired them.

The complete 70-episode script set. Each script in its final approved version. File name: EPISODE-NUMBER_EPISODE-TITLE_APPROVED-DATE.docx. Every script in the set, not only the paywall episode and the key structural episodes.

The direction brief templates for each scene type. File name: SCENE-TYPE_DIRECTION-BRIEF-TEMPLATE.docx. The template that the sequel's creative director will use to brief generation operators on each scene type.

Where to store it:

A creative documents folder with sub-folders for arc map, character bibles, style guide, scripts, and direction brief templates. Version history enabled for all documents. The creative documents folder is the archive that the sequel's creative director works from during brief development.

Category 4: The Production Operations Archive

The production operations archive preserves the operational documentation that allows a new team member joining the sequel production to understand the primary series' production process without requiring briefing from the primary series' team members.

What to archive:

The complete generation log in the format described in the production management stack post. Not a summary. Every session, every output, every quality criteria result, every revision count, every approval status. The generation log allows the sequel's production coordinator to identify which character configurations produced the most consistent results and which produced the most revisions, informing the sequel's pre-production decisions.

The production schedule from the primary series. The week-by-week timeline that produced the eight-week delivery. The sequel's production coordinator uses this as the baseline for the sequel's schedule, adjusting only the stages that the sequel's specific scope requires.

The quality review notes from the creative director's batch reviews. Every note the creative director provided on generation outputs across the primary series' production. These notes identify the recurring quality failures that the sequel's generation operators should be briefed to avoid before production begins.

The delivery package index from the primary series. The complete list of files delivered to the platform, with file names, specifications, and the delivery date. The sequel's post-production specialist uses this as the confirmed delivery specification for the sequel, verifying with the platform whether the specifications have changed.

Where to store it:

A production operations folder with sub-folders for the generation log, the production schedule, the quality review notes, and the delivery documentation. The production operations folder does not need version history for all documents, but the generation log must be version-controlled because it is the record of what was produced across the full primary series.

The Archive Handoff Protocol

The archive is not complete when it is assembled. It is complete when a new team member who was not on the primary series can locate every file they need for the sequel's pre-production without asking anyone from the primary series team.

The archive handoff protocol tests this: at primary series delivery, the production coordinator assigns a team member who was not on the primary production to locate three specific items from the archive using only the archive's folder structure and file naming conventions — no verbal direction permitted. The three items: the session-open consistency reference frame for the controlled alpha character, the ControlNet geometry reference for the authority close-up camera position, and the approved paywall episode arc map specification.

If the team member locates all three items within ten minutes without assistance, the archive passes the handoff protocol. If any item cannot be located within ten minutes, the archive has a naming, structure, or documentation failure that must be corrected before the archive is considered complete.

The handoff protocol is a five-minute test that prevents an 18-month archive failure.

What the Sequel Team Will Actually Need

At 18 months post-delivery, when the sequel commission arrives, the production team will need exactly five things from the archive in the first week of pre-production:

The session-open consistency reference frames for each primary character. These frames are the first files the generation operators load in the sequel's first production session.

The Soul ID model identifiers or character model parameters. Without these, the character models cannot be reloaded on the current generation tool's infrastructure.

The arc map. The sequel's arc map is built in direct continuation of the primary series' arc, which requires the primary series' arc map as the structural foundation.

The style guide at generation specification level. The sequel's visual register must match the primary series without requiring a rebuild from zero.

The ControlNet geometry reference library. The sequel's camera positions are the primary series' camera positions, continued.

If all five are immediately accessible from the archive on day one of the sequel's pre-production, the archive has fulfilled its commercial function. The sequel begins from the primary series' established infrastructure rather than from a rebuild.

Axis AI Studios Perspective

The archive is not a post-production administrative task. It is the sequel's pre-production starting point, assembled at primary series delivery. At Axis AI Studios, the archive structure described in this post is built into the production workflow from week one of every commission. The naming conventions are specified before the first generation session. The folder structure is established before the first reference image is produced. The generation log template is active from the first session.

The archive that is built correctly from the beginning of the production is the archive that enables the sequel at fraction of the rebuild cost. The archive that is assembled after delivery from whatever was saved during production is the archive that produces rebuild costs that the sequel commission budget did not anticipate.

For businesses commissioning AI-native vertical drama with franchise ambitions, the archive structure is a pre-production brief question: what naming conventions, folder structures, and documentation standards will the production partner use to ensure the sequel's pre-production can begin from established infrastructure 18 months from now?

Reach out at business@axisaistudios.com for commissioning conversations or for production team members who want to understand archive discipline at professional production standard.

FAQ

How Long Does Archive Assembly Take at Primary Series Delivery?

With the naming conventions and folder structure established at pre-production, archive assembly at delivery takes one to two hours. The production coordinator populates the delivery archive folders from the materials that have been correctly named and stored throughout the production. Without established conventions, archive assembly at delivery takes one to two days as the team reconstructs naming conventions and searches for files that were not stored with consistent naming.

What Cloud Storage Service Is Best for the Production Archive?

Google Drive and Dropbox are both viable for the production archive. The critical requirement is version history: both services provide version history that allows restoration of previous file versions if an archive file is accidentally modified or deleted. The secondary requirement is access management: the archive should be accessible to the commissioning party, the production company, and any future production team that works on sequel commissions, with clear access permissions for each group.

What Happens to the Archive if the Production Partner Relationship Ends Before the Sequel?

The commissioning party who owns the IP also owns the right to access the production archive materials. The production agreement should specify that the production company will provide a complete archive handoff to the commissioning party within 30 days of delivery. This provision ensures that if the production partner relationship ends before the sequel commission, the commissioning party has the archive materials needed to brief a new production partner for the sequel without rebuilding from zero.

Further Reading

For the character bible that the character archive described in this post preserves, the guide to how to build a vertical drama character bible for AI generation covers the complete character bible structure at the specification level that the archive must preserve.

For the sequel commission that the archive enables, the guide to when to commission a sequel and when to start a new franchise covers the performance metrics that drive the sequel decision and the production infrastructure that the archive enables.

For the character asset library that the archive feeds into across multiple franchise series, the guide to building an AI character asset library from one series to a franchise covers Soul ID training, character model management, and franchise consistency infrastructure.