ButterCutButterCut

Netflix, Prime Video and Disney+ Subtitle Specs: The Compliance Guide for Indian Content Teams (2026)

Apr 27, 20268 min readBy ButterCut Team

The technical specifications Netflix, Prime Video and Disney+ actually enforce on subtitle files, and where Indic scripts and Hinglish dialogue break compliance in ways Western guides never cover.

Editorial illustration of a towering fortress gate built from rulers, calipers and stopwatch dials, with cracked rejected subtitle bars lying at its base and a single orange subtitle bar passing through a narrow slot of warm light
OTT spec validation is a gate built from measurements; most files crack against it, and only spec-compliant ones pass on the first attempt.

Your subtitle file can be perfectly translated, perfectly timed, culturally flawless, and still get rejected in seconds. Not by a person. By a validator that checked the file format, counted characters per second, and bounced it before any human at the platform pressed play.

That's how OTT subtitle delivery works now. Netflix, Prime Video, and Disney+ each publish detailed technical specifications, run automated QC against them, and treat non-compliance as an automatic redelivery. For an Indian content team pushing regional-language files, every rejection is days added to a release date. This guide covers what the platforms actually enforce, and where Indic-language content trips the rules in ways the Western-focused guides never mention.

OTT subtitle specifications are the technical and editorial rules a streaming platform enforces on subtitle files, covering file format, reading speed, timing, line length, and permitted characters. They work by running every submitted file through automated validation before human review, rejecting non-compliant files outright. Most commonly used for delivering content to Netflix, Prime Video, and Disney+ Hotstar.

The specs all three platforms enforce

Research from Ekitai Solutions comparing all three platforms' published documentation found they align on the core limits: a maximum of two lines per subtitle and 42 characters per line. For your team, this means one master set of formatting rules covers most of your output, and the platform differences live in file formats and edge cases.

RuleNetflixPrime VideoDisney+
File formatTTML1 (.xml/.ttml); IMSC 1.1 for JapanesePer Video Central documentation; TTML-familyIMSC 1.1 primary
Max lines per subtitle222
Max characters per line424242
Reading speed17 CPS adult, 13 CPS childrenAligned industry limitsAligned industry limits
Min/max event duration5/6 second to 7 secondsComparable limitsComparable limits
Gap between subtitlesMinimum 2 framesComparableComparable
Forced narrativesSeparate FN handling per TTSGSeparate FN file per dubbed audio languagePer style guide via approved vendors

Netflix: the deepest rulebook

Netflix's Timed Text Style Guide is the most detailed in the industry and functions as the de facto standard everyone else references. The rules that catch teams most often:

  • Timing: every subtitle event must stay on screen at least 5/6 of a second (20 frames at 24fps) and at most 7 seconds, with a minimum 2-frame gap between consecutive events. Files that violate these get flagged automatically.
  • Reading speed: Languages Unlimited's breakdown of the guide puts the ceiling at 17 characters per second for adult content and 13 for children's content. If a line exceeds the limit, the fix is splitting the subtitle, not shortening its display time.
  • Glyph list: only characters on the Netflix Glyph List may appear in a file. This matters enormously for Indic scripts, covered below.
  • KNP tables: for episodic content, Key Name and Phrase tables are mandatory so character names, title cards, and recurring terminology stay consistent across every episode and season, and they must be delivered to Netflix for QC.
  • Timestamps: all files start at hour 00. Tape-style hour 01 or hour 10 starts are rejected.
  • No currency conversion: money mentioned in dialogue stays in the original currency in every language's file.

Netflix also strongly encourages producers to route delivery through Netflix Preferred Vendors, whose entire value is checking files against current specs before submission. That tells you how the platform itself rates the odds of a first-time team passing validation unassisted.

Prime Video and Disney+: where they differ

Prime Video's sharpest edge is forced narratives. Amazon requires a separate Forced Narrative timed text file for every dubbed audio language in a multi-track package, and the language-locale tag of that FN file must exactly match the video mezzanine. A mismatch doesn't produce an error you can read; it produces forced narratives that silently fail to display when subtitles are off. Teams often discover this after launch.

Disney+ runs on IMSC 1.1 as its primary format and manages its style guide through approved vendors rather than fully public documentation, which in practice means you learn the rules through rejections unless your vendor already knows them.

The stakes in India are not niche. Research from Ekitai Solutions notes that Disney+ Hotstar reports nearly 40% of its India content viewership coming from regional languages. For an acquisition or localization head, this means your Tamil, Telugu, and Bengali subtitle files aren't a compliance afterthought; they carry close to half your audience.

Where Indic languages make compliance harder

The published guides are written for teams delivering French and Spanish. Indic delivery adds failure modes they never cover:

  • Glyph list collisions: Devanagari, Tamil, Telugu, and Bengali scripts use conjunct consonants and combining marks that must render within the permitted character set. A transliteration choice that looks fine in a subtitle editor can fail glyph validation.
  • Reading speed vs script density: 42 characters of Hindi carries different information density than 42 characters of English, and Indic scripts hit CPS ceilings differently. Condensing without losing meaning is a per-language editorial skill, applied thousands of times per file.
  • Code-switched dialogue: Hinglish dialogue forces constant judgment calls about what counts as foreign dialogue requiring treatment under the style guide versus natural speech to subtitle as-is. Netflix's own guidance requires accurate, authentic transcription of non-standard speech, which generic pipelines and offshore human teams both routinely flatten.
  • Song lyrics: Indian content is dense with music, and lyric subtitling rights on OTT platforms are a rights minefield with platform-specific rules that need flagging per title, not a blanket policy.

Manual QC vs encoded compliance

Here's the structural problem. Every rule above is checkable, which means every rule is automatable. Yet most vendors comply the manual way: a human subtitler works, a human QC reviewer checks timing, CPS, line length, and glyphs against a checklist, finds violations, and sends the file back around the loop. Each loop adds days, and the checklist grows with every platform update.

The alternative is encoding the spec into the production system itself, so output is generated compliant rather than corrected into compliance. That's how ButterCut's subtitle pipeline handles platform specs: Netflix timing rules, CPS ceilings, line limits, and format requirements are constraints the engine writes within, not errors a reviewer hunts afterward. The same mechanism handles custom rules too. If your brand style guide says numerals under ten are spelled out in Hindi but not in Tamil, that gets encoded once and enforced on every file, instead of living in a PDF your fifth freelancer never read.

And because the pipeline learns from corrections, a spec update or a new client rule propagates immediately across all languages, rather than through a re-briefing cycle with a translator bench.

Where it works

  • Recurring episodic or catalog delivery to OTT platforms, where KNP consistency and repeat compliance matter most
  • Multi-platform releases, where one source must produce Netflix TTML and Disney+ IMSC outputs without duplicate work
  • Regional-language catalogs at Hotstar-style scale, where 40% of viewership rides on Indic file quality
  • Teams burned by redelivery cycles pushing back release dates

Where it doesn't

  • A single festival film heading to Netflix as an Original, where an NPV relationship and a senior human subtitler are worth the premium
  • Languages or scripts outside the pipeline's trained set
  • Titles with heavy licensed-music lyric subtitling, where rights clearance is a legal workflow no production system solves

FAQ

What subtitle format does Netflix require?

TTML1, delivered as .xml or .ttml files, for all languages except Japanese, which requires IMSC 1.1. Positional data must use percentage values, not pixels. SRT files are not accepted for Netflix delivery, and non-compliant formats are rejected before human review.

What is the maximum reading speed for Netflix subtitles?

Up to 17 characters per second for adult content and 13 for children's content in most languages. Each subtitle event must display between 5/6 of a second and 7 seconds, with at least a 2-frame gap between consecutive events.

Why do OTT platforms reject subtitle files?

Most rejections are automated: wrong file format, reading-speed violations, missing frame gaps, characters outside the glyph list, or timestamps starting at the wrong hour. Linguistic quality is judged later; technical validation happens first and bounces files instantly.

Do Prime Video and Disney+ use the same specs as Netflix?

The core limits align: two lines maximum and 42 characters per line across all three. Formats differ: Netflix requires TTML1, Disney+ uses IMSC 1.1, and Prime Video adds strict forced-narrative file requirements per dubbed audio language.

Netflix, Prime Video, and Disney+ enforce subtitle specifications through automated validation: TTML or IMSC file formats, two-line and 42-character limits, 17 CPS reading speed, and strict timing rules, with non-compliant files rejected before human review. For Indian content teams, Indic scripts and code-switched dialogue add failure modes Western guides skip. Encoding these specs into the production pipeline, as ButterCut does, replaces manual QC loops with first-pass compliant delivery.

If your last OTT delivery bounced and pushed a release date, the fix isn't a longer QC checklist. Send ButterCut one episode and your target platform, and get back spec-compliant files in every Indic language, validated before you submit, not after you're rejected.

Sources