How to Import and Organize a Short Drama Library

Short drama catalog management comes down to one habit: decide the structure before you upload anything. Fix how a story becomes a title, how a title becomes numbered episodes, where the trailer sits, which subtitle files attach to which episode and what each title's territory and dates are. Then prepare files, import in batches, check each batch and only then open the shelves to viewers.

This is a runbook for an operator or content manager with dozens or hundreds of titles. It follows the way a library-first white-label MoboReels clone organizes content: categories, titles typed as movie or web series, and numbered episodes, each with video, artwork, an optional subtitle file, a coin price and a lock. It does not repeat where to source titles (see where to get micro drama content) or the rights terms (see what a micro drama licensing deal should cover).

How a library is structured

Three levels carry everything: a category (a genre shelf such as romance or revenge), a title (one story, typed as a movie or a web series) and an episode (one numbered video). Each level has its own switches and artwork. A title has a poster and a banner, a plot text, a territory list and dates. An episode has a video file, a duration, an optional subtitle file, a coin price and a lock.

The trailer sits inside the title as episode zero. That is a useful rule because promo material follows the same tools as any episode: it is published, ordered, targeted by region and hidden in the same way. A viewer sees the trailer in a discovery view, and an editor manages it without a separate screen.

Because the price and lock sit on the episode, you can price unevenly. An opener can be free, a middle episode cheap and the cliffhanger priced properly. That choice belongs in your metadata sheet from the start, not in a later edit. For the logic behind it, see episode paywall strategy and free episodes.

Prepare files before import

Most import problems are file problems. Work through the checklist per title before it reaches the console.

• One master per episode. A single file with a stable name; no "final_v2_REAL" copies left in the folder.
• Consistent vertical format. Same aspect ratio, resolution and codec across the library, so the player behaves the same on every title.
• Burned-in text decided. Know which episodes carry hard subtitles from the source. Do not add soft subtitles in the same language on top of them.
• Duration recorded. Read the real duration from the file, not from a supplier's sheet that may be wrong.
• Artwork set. A poster and a banner per title at the sizes your design uses, plus an optional image per episode.
• Subtitle files named to match. One file per episode and language, UTF-8.
• Trailer cut. A short clip, ready to be episode zero.
• Rights data to hand. Territory list, start and end dates, and monetization allowed, taken from the signed contract.

A naming rule that survives hundreds of titles

Pick a pattern and never break it. A workable one is series-slug_e001.mp4 for the video, series-slug_e001.en.vtt for the English subtitle and series-slug_poster.jpg for artwork. Three digits for the episode keeps files sorted correctly up to 999 episodes. Use lowercase, hyphens, no spaces and no accents. When a supplier's names differ, rename before import; scripts that match files by name will fail on exceptions.

The metadata sheet

A spreadsheet is the backbone of a clean import. One row per title, and a second sheet with one row per episode. The title sheet below shows the fields to settle before you begin.

The episode sheet carries the title key, the episode number, the file name, the duration, the subtitle languages and the lock and price. Validate it with simple checks in the spreadsheet: no duplicate episode numbers inside a title, no gaps in numbering, every file name present on disk, every price filled where the lock is on. Fifteen minutes of formulas saves a day of manual repair.

The bulk import path

Work in batches. The aim is to find a mistake in a rule after twenty titles, not after three hundred.

1. Pilot with two titles. One series, one movie, from the real files. Walk them from import to the viewer app on web and phone.
2. Fix the rules. Whatever broke in the pilot (naming, durations, artwork size, subtitle encoding) is fixed in the sheet and the scripts.
3. Batch one: twenty titles. Choose one genre or one licensor. Load them, then run the QA checklist below.
4. Review and correct. Record every defect with its cause. A repeated cause means a rule change, not twenty manual fixes.
5. Scale in larger batches. Increase the batch size while the defect rate stays low.
6. Set switches last. Keep new titles hidden until QA passes, then turn on the active flag and the shelf flags.
7. Freeze and log. After the final batch, export the sheet as the record of what was loaded and when.

Publishing a title through the console means uploading video, artwork and subtitles and setting prices and locks. For a large catalog, we set up a bulk import routine for your build, working from your metadata sheet, so editors do not key hundreds of records by hand. We agree batch sizes and the mapping from your sheet to categories, titles and episodes at kickoff; contact us through the contact page before you prepare the sheet, so it matches. If you license titles from partners that expose an interface, we can fill the library from it on a schedule. See the MoboReels clone features for the catalog tools behind this.

Every upload is read at the byte level before it is accepted, and only image, video and subtitle families pass. A renamed executable or archive is discarded. That check protects the library from a careless supplier, but it also means a mislabeled file will be rejected, so keep your extensions honest.

Subtitle formats and audio tracks

The player accepts WebVTT and SRT files. The language and format are recorded for each episode, and the file appears as a subtitle track on web and mobile. Pick one format for the library and convert the rest.

The W3C WebVTT specification says a file must start with the WEBVTT line, use UTF-8, list cues in order of start time and use timestamps with three-digit milliseconds. It also says a file holds one kind of content. If a supplier's file fails to load, check those four points first. A frequent defect is a comma where a period belongs after conversion from SRT.

Subtitle QA

• The file name ends in the right language code and attaches to the right episode.
• Timing matches the final cut; subtitles from a first cut drift against a re-edited episode.
• Lines are short enough for a vertical screen; two lines at most.
• Right-to-left languages display in the right direction. The W3C internationalization guidance explains why direction must be set explicitly and not guessed.
• Character encoding is UTF-8 and special characters display correctly.

Audio tracks work like subtitles at the episode level: an episode can carry selectable audio tracks that play from stored files, which suits dubbed versions. For the choice between dubbing and subtitling, see dubbing or subtitles for short drama.

Seasons and numbering

A title holds numbered episodes. A story with more than one season is normally modeled as separate titles, one per season, with a clear naming rule such as "Series Name, Season 2". Link them for viewers with artwork, plot text and a shelf. If your catalog depends on a season level above the episode, we add it to your build; say so at planning so numbering, locks and revenue reports are designed around it.

Numbering needs three rules.

1. Episode one is the first story episode. The trailer is episode zero.
2. Numbers are continuous, with no gaps. A gap shows as a missing episode to the viewer.
3. Numbers are never reused for different content. If an episode is removed, renumber the run, do not slot a new one into the old number.

When you drag an episode to a new position, the title renumbers and the locks are reapplied afterward. That keeps the numbering consistent across the web app, the mobile app and the console, so re-cutting a series after launch does not need manual data repair.

Reordering without hurting paid viewers

Before launch, reorder freely. After launch, treat the order as a promise. A viewer who spent coins on episode twelve expects to find the same story under that number tomorrow. Reordering moves content under existing numbers, and because locks are reapplied after a move, a changed order can change which episodes are locked.

Use these rules after launch.

• Append, do not insert. Add new episodes at the end of the run.
• Fix errors early. If an episode is wrong, replace its file in place, so the number and the unlock record stay put.
• Reprice by band, not by shuffle. Change the coin price on an episode instead of moving it.
• Test on a hidden copy. Try a re-cut on a hidden test title first and check which episodes are locked afterward.
• Tell viewers. A note on the title page explains an update to a series people are mid-way through.

How unlocks are keyed to episodes in your build is something we confirm with you at kickoff, since it decides what a viewer who paid for episode twelve keeps after a re-cut.

Several brands from one library

Operators with more than one audience often run several brands over the same titles: one app for one language market and another for a different one. Each brand runs as its own service with its own database, domains and store listing. The library work is done once, in the files and the sheet; it is loaded into each service.

Design the sheet for that from day one. Add a column per brand for visibility, price band and territories. Keep file names brand-neutral, and keep brand-specific artwork in a separate folder. Remember that a license names the operator and the brand, so check that a licensor allows the second brand before you load their titles into it. Rights follow the contract, not the file.

Discovery without a recommendation engine

A library of hundreds is useless if viewers cannot find the next title. Rails are composed by your editors from the catalog: a banner carousel, trending, new releases, category shelves and a coming soon shelf. Search reads title and plot text and can narrow by genre. Viewers save a series to My List, resume where they stopped and set a reminder on an upcoming title. An unreleased title shows on its own shelf with an announce date and stays off every other rail.

The weekly job is editorial. Someone owns the home screen: what sits under the banner, which titles trend, what is coming next. Series and episode reports tell them where viewers paid and where they stopped. Rules of thumb:

• Lead every shelf with a title that has free episodes, so a new viewer can start without paying.
• Rotate the banner weekly; stale banners tell viewers nobody is looking after the app.
• Use tags and category consistently, because search and shelves only work if the metadata does.
• Promote back-catalog titles that earn with rewarded ads or a VIP pass, using the revenue-by-series report as evidence.

QA checklist for each batch

Run this list on a sample of each batch, and on every title in the first batch.

• The title appears in the right category and in the right type.
• The poster and banner display without cropping on web and phone.
• Episode count matches the sheet; numbers run from one without gaps; the trailer is episode zero.
• Every episode plays from the first frame to the last, with sound.
• Durations shown match the file.
• Locks and coin prices match the price plan; free episodes are free.
• Subtitle tracks load, are in the right language and match the timing.
• Audio tracks, where present, switch correctly.
• Territory chips and dates match the contract.
• A viewer in an excluded country does not see the title in lists or search.
• Search finds the title by name and by a word from the plot.
• Hidden titles stay hidden until you switch them on.

Common import errors and their causes

A worked example with invented numbers

Say a catalog has 300 titles of 60 episodes each. That is 18,000 episode files and 18,000 or more subtitle files if you carry one language. Checking every file by eye is not realistic. Instead, the content manager checks the first batch of 20 titles fully, which is 1,200 episodes, and 10% of the episodes in every later batch, chosen at random. If the sample defect rate stays under 1%, the batch passes. If it exceeds 1%, the batch is held and the cause is found. These thresholds are an example to adapt, not an industry standard. The point is to agree a sampling rule in advance so QA is a decision and not a mood.

Protect the library as it grows

A library holds other people's property, so organize it with protection in mind. Keep masters apart from playback files, grant upload access to the few staff who need it and keep a log of who changed what. Under U.S. law, as the Copyright Office's Circular 1 explains, a copyright owner holds the exclusive rights to reproduce and distribute a work, so an import mistake that publishes a title in the wrong territory or after its end date is not just untidy; it can be a breach. Territory and dates in the sheet are the control. For file-level protection, such as expiring links with country rules described in Cloudflare Stream's documentation, see protecting a licensed drama catalog from piracy.

What to do next

Write the naming rule and the metadata sheet this week, and test them on two real titles. Settle the questions that change the structure before the first large batch: whether you need a season level, whether you add audio tracks, which brands share the library and whether you need adaptive delivery. Then plan the import in batches with a QA gate between each. For the cost and scope of preparing and loading a large catalog, read the MoboReels clone development cost page, then tell us the size and state of your library so we can plan the import for your build on a ready-made MoboReels clone.