> ## Documentation Index > Fetch the complete documentation index at: https://docs.decktalk.app/llms.txt > Use this file to discover all available pages before exploring further. # Build artifacts > The shape of every file DeckTalk writes under build/, including the words files, manifest, timeline, beats, recording sidecars, section mp4s, the final video, captions, and chapters. These files are part of the public contract. The page runtime reads `beats.json`, and your own scripts may read any of them. Every JSON file is written atomically, so a reader never sees a partial file. Field names in the JSON match the dataclasses in `decktalk.artifacts`. ```text theme={null} build/ audio/ NN-slug.mp3, NN-slug.words.json, manifest.json, narration.mp3, timeline.json, beats.json rec/ NN-scene.webm, NN-scene.json out/ NN-section.mp4, .mp4, -YYYYMMDD-HHMM.mp4, .srt, .vtt, .chapters.txt, slates/ shots/ step-.png, section-NN-at-s.png sfx/ ambience.mp3, .mp3, and a .manifest.json beside each music/ underscore.mp3, its parts, and its manifest ``` ## `build/audio/NN-slug.words.json` A list of words with times within that section's mp3. ```json theme={null} [ { "word": "Welcome", "start": 0.7, "end": 1.113 }, { "word": "This", "start": 1.133, "end": 1.547 } ] ``` The first word of the first spoken section starts after the lead break. Punctuation is stripped from each word. ## `build/audio/manifest.json` One entry per narrated section. The `hash` is what the cache compares. | Field | Meaning | | --------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- | | `script`, `model`, `output_format` | What produced the audio. A silent build records `silent-placeholder` as the model. | | `estimated`, `estimate_basis` | True after a silent build, with the pacing used, such as `150 wpm + declared pauses`. | | `total_seconds` | The sum of the segments. | | `segments.NN.index`, `.title` | The section. | | `segments.NN.file`, `.words_file` | The mp3 and the words file. | | `segments.NN.hash` | The provider, voice, settings, and text hash, or `silent`. | | `segments.NN.words`, `.est_seconds`, `.duration_seconds` | The word count, the pacing estimate, and the real length. | | `segments.NN.target_seconds`, `.speech_end_seconds`, `.tail_padded_seconds` | The time budget from the heading, where speech ends, and how much silence was added. | ## `build/audio/timeline.json` Section and word times, absolute within `narration.mp3`. `speech_end` is where the last word ends, and the section runs on to `end` in silence. ```json theme={null} { "narration": "narration.mp3", "total_seconds": 111.9, "sections": { "01": { "title": "Open", "start": 0.0, "end": 10.15, "duration": 10.15, "speech_end": 9.78, "words": [ { "word": "Welcome", "start": 0.7, "end": 1.113 } ] } }, "estimated": true } ``` ## `build/audio/beats.json` The resolved cues per section, as the string the page receives in `?beats=`. The seconds are relative to the section's start, and a section with no resolved cue is left out. ```json theme={null} { "01": "1.1a@0.0,1.1b@2.43,1.1c@6.77", "02": "2.1a@2.3,2.1b@5.51,2.1c@9.19,2.1d@12.41,2.2@13.79,2.2edit@17.92" } ``` ## `build/rec/NN-scene.json` The sidecar of a recording: what the recorder did for one section, and where narration t=0 sits in the webm. ```json theme={null} { "url": "file:///home/me/my-lesson/deck/index.html?scene=1&beats=1.1a%400.0%2C1.1b%402.43%2C1.1c%406.77&t0=signal", "requested_seconds": 10.450000000000001, "settle_seconds": 1.467, "load_seconds": 0.038, "lead_seconds": 1.504, "lead_in_seconds": 1.44, "lead_method": "cover (36 magenta frames)", "warnings": [] } ``` | Field | Meaning | | ------------------------------ | ---------------------------------------------------------------------------------------------------------------------------- | | `url` | The page URL the recorder opened, including `scene`, `beats`, and `t0=signal`. | | `requested_seconds` | How long the recording was asked to run. | | `settle_seconds` | How long the recorder waited after load before starting the clock. | | `load_seconds`, `lead_seconds` | Wall-clock estimates from the recorder. | | `lead_in_seconds` | The first clean frame after the magenta cover, written by `measure`. The assembler trims here. | | `lead_method` | How `measure` found it. A value beginning `NO COVER` is a guess. | | `warnings` | `window.__decktalk.warnings` read from the page after the recording, such as an unknown cue id or a KaTeX that never loaded. | ## `build/out/` | File | Meaning | | -------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `NN-section.mp4` | Each section cut to its span, at the project's frame size, video only. | | `.mp4` | The video, with a chapter per section muxed in. | | `-YYYYMMDD-HHMM.mp4` | A timestamped copy of the same video. | | `.srt`, `.vtt` | Captions from the word timestamps, absolute in the mp4, with the script's punctuation and case. A caption holds at most two lines of 42 characters, ends at a sentence, a pause longer than 1.5 seconds, or seven seconds, and never spans a section. | | `.chapters.txt` | An ffmetadata file with one `[CHAPTER]` per section, titled from `decktalk.toml`, which is what the mp4 carries. | | `slates/NN-slate.png` | Rendered slates for clip sections whose file is missing. | ```text build/out/my-lesson.srt theme={null} 1 00:00:00,700 --> 00:00:06,767 Welcome. This is a narrated lesson, cut to the word. Every visual you see 2 00:00:06,767 --> 00:00:09,980 lands on the word that introduces it. ``` ```text build/out/my-lesson.chapters.txt theme={null} ;FFMETADATA1 [CHAPTER] TIMEBASE=1/1000 START=0 END=10160 title=Open ``` ## `build/shots/` `step-.png` for each step, in a subdirectory named after the page when the project has several pages, and `section-NN-at-s.png` for each frame taken from a playing section.