Add subtitles to videos

🤖/video/subtitle adds subtitles and closed captions to videos.

Usage example

{
  "steps": {
    "captioned": {
      "ffmpeg_stack": "v7",
      "preset": "empty",
      "robot": "/video/subtitle",
      "style": "social",
      "subtitles_type": "burned",
      "use": {
        "steps": [
          {
            "as": "video",
            "name": ":original"
          },
          {
            "as": "subtitles",
            "name": "transcript"
          }
        ]
      }
    },
    "transcript": {
      "format": "json",
      "provider": "aws",
      "robot": "/speech/transcribe",
      "source_language": "en-US",
      "use": ":original"
    }
  }
}

Parameters

• interpolate

  boolean | Record<string, boolean>

  Controls whether Assembly Variables are interpolated for individual instruction fields.

  By default, most Robot instruction fields interpolate Assembly Variables. Set this to false to treat every instruction field as literal text, or set an individual field path to false to treat only that field as literal text. For Robot-specific fields that are literal by default, set this to true or set that field path to true to opt back into interpolation.

  Use field names such as path, or dotted paths such as ffmpeg.vf for nested objects.

• output_meta

  Record<string, boolean> | boolean | Array<string>

  Allows you to specify a set of metadata that is more expensive on CPU power to calculate, and thus is disabled by default to keep your Assemblies processing fast.

  For images, you can add "has_transparency": true in this object to extract if the image contains transparent parts and "dominant_colors": true to extract an array of hexadecimal color codes from the image.

  For images, you can also add "blurhash": true to extract a BlurHash⁠ string — a compact representation of a placeholder for the image, useful for showing a blurred preview while the full image loads.

  For images, "thumbhash": true instead extracts a base64-encoded ThumbHash⁠ into meta.thumbhash, together with meta.has_alpha (whether an alpha channel exists, even when fully opaque). It describes EXIF-oriented pixels and uses the first frame of animated images. Extraction is best-effort: images above 40 megapixels, unsupported formats, or a failed/bounded decode produce no placeholder. Successful extraction adds a metadata charge equivalent to 20% of that file's bytes. No ThumbHash surcharge applies when disabled or when no hash is produced.

  Set this option on the Step producing the image, such as /upload/handle for uploaded originals or /image/resize for processed outputs. Setting it only on /transloadit/store does not request extraction: storage preserves the producer's metadata. Transloadit Storage persists a generated hash with its immutable version and returns it in stored results and native asset reads. Direct S3 uploads do not generate placeholders. A placeholder contains image information, so protect it with the same access controls as the full image.

  For videos, you can add the "colorspace": true parameter to extract the colorspace of the output video.

  For videos, you can also add "interlaced": true to detect whether the video is interlaced. This combines the cheap ffprobe field_order flag with a bounded idet sampling pass over the first frames of the source, exposing interlaced, field_order, and a diagnostic interlace_detection object under file.meta. This is computationally expensive and billed accordingly.

  For audio, you can add "mean_volume": true to get a single value representing the mean average volume of the audio file.

  You can also set this to false to skip metadata extraction and speed up transcoding.

• user_meta

  Record<string, any>(default: {})

  Adds custom JSON metadata to each emitted file without changing its contents. Nested objects and arrays are supported.

  Inheritance depends on the Robot. Values merge with existing user_meta on the output file; the current Step replaces matching top-level keys. Assign required keys explicitly when a Robot creates fresh outputs.

  In processing Steps, ${file.*} refers to the first input and ${result.*} to the emitted file. Values are evaluated per output after the Robot runs, before subsequent metadata extraction and temporary storage. On :original, values are evaluated per upload before metadata extraction.

  Downstream Steps read ${file.user_meta.key}. See Custom metadata for a complete example and inheritance rules.

• result

  boolean(default: false)

  Whether the results of this Step should be present in the Assembly Status JSON

• queue

  batch

  Setting the queue to 'batch', manually downgrades the priority of jobs for this step to avoid consuming Priority job slots for jobs that don't need zero queue waiting times

• force_accept

  boolean(default: false)

  Force a Robot to accept a file type it would have ignored.

  By default, Robots ignore files they are not familiar with. 🤖/video/encode, for example, will happily ignore input images.

  With the force_accept parameter set to true, you can force Robots to accept all files thrown at them. This will typically lead to errors and should only be used for debugging or combatting edge cases.

• ignore_errors

  boolean | Array<meta | execute>(default: [])

  Ignore errors during specific phases of processing.

  Setting this to ["meta"] will cause the Robot to ignore errors during metadata extraction.

  Setting this to ["execute"] will cause the Robot to ignore errors during the main execution phase.

  Setting this to true is equivalent to ["meta", "execute"] and will ignore errors in both phases.

• use

  string | Array<string> | Array<object> | object

  Specifies which Step(s) to use as input.

  • You can pick any names for Steps except ":original" (reserved for user uploads handled by Transloadit)
  • You can provide several Steps as input with arrays:

    {
      "use": [
        ":original",
        "encoded",
        "resized"
      ]
    }
    

  • You can also tag input Steps with as to pass semantic intent to robots:

    {
      "use": [
        {
          "name": ":original",
          "as": "image"
        },
        {
          "name": ":original",
          "as": "mask"
        }
      ]
    }
    

  Tip

  That's likely all you need to know about use, but you can view Advanced use cases.

• ffmpeg

  object

  A parameter object to be passed to FFmpeg. If a preset is used, the options specified are merged on top of the ones from the preset. For available options, see the FFmpeg documentation⁠. Options specified here take precedence over the preset options.

• ffmpeg_stack

  v6 | v7 | v8 | string(default: "v6.0.0")

  Selects the FFmpeg stack version to use for encoding. We currently recommend using "v7". The exact versions "v6.0.0", "v7.0.0", and "v8.0.0" are legacy values that remain accepted for backward compatibility. Deprecated "v5.x" values are also accepted.

• width

  string | number | null

  Width of the new video, in pixels.

  If the value is not specified and the preset parameter is available, the preset's supplied width will be implemented.

• height

  string | number | null

  Height of the new video, in pixels.

  If the value is not specified and the preset parameter is available, the preset's supplied height will be implemented.

• preset

  android | android-high | android-low | android_high | android_low | dash-1080p-video | dash-1080p_video | 195 more ↓

  Converts a video according to pre-configured settings.

  You can use the value 'empty' here if you specify your own FFmpeg parameters using the Robot or do not want Transloadit to set any encoding settings.

• subtitles_type

  burned | external | burn(default: "external")

  Determines if subtitles are added as a separate stream to the video (value "external") that then can be switched on and off in your video player, or if they should be burned directly into the video (value "burned" or "burn") so that they become part of the video stream.

• style

  social

  Applies a ready-made appearance for burned subtitles: heavy Poppins text, a thick outline, and a soft black shadow. JSON captions render in uppercase in the lower third. By default, JSON font sizing uses the shorter side of the video and fits each caption within the safe width, preferring one line and using at most two. Words are never split. Each caption grows from 85% to full size over 120 milliseconds, then its words stay in place while the highlight changes color. The transcript file is preserved. Existing styling parameters override the preset; an explicit font_size disables automatic fitting. ASS files retain their own styles. With timestamped JSON from /speech/transcribe, the active word is highlighted until the next word or caption starts; other words use font_color. Trailing commas and periods are omitted from social JSON captions; question marks and exclamation marks are preserved. Word highlighting and grouping require JSON input.

• highlight_color

  string | string(default: "FFD60A")

  The accent color of the active word in JSON subtitles. Only applies to burned JSON subtitles. With "style": "social", the previous word stays highlighted through short gaps until the next word or caption starts. Without the social style, gaps use font_color for the entire caption.

  Use "rgb", "rrggbb", or "aarrggbb" format, optionally prefixed with #. Three-digit colors are expanded to six digits. Colors without alpha are fully opaque. For eight-digit colors, the first two digits specify transparency ("00" is fully opaque, "ff" is fully transparent).

• max_words_per_caption

  string | number(default: 3)

  The target maximum number of timestamped JSON words shown in a caption. Captions break earlier at sentence-ending punctuation or pauses longer than 0.4 seconds. With "style": "social", a trailing English function word such as “by,” “to,” or “the” moves to the next caption when possible. Short phrases such as “line of code” can stay together, and a sentence-final single word can join the previous caption. These adjustments allow at most one extra word per caption, without crossing sentence endings or long pauses. Single-word captions are preserved when the limit is 1. A caption stays visible until the next caption starts when the pause is at most 0.4 seconds, and disappears at its last word’s end during longer pauses. Does not change SRT, VTT, or ASS cues.

• border_style

  box | outline | shadow(default: "outline")

  Specifies the style of the subtitle. Use the border_color parameter to specify the color of the border.

• border_color

  string | string(default: "000000")

  The color for the subtitle border. The default is opaque black. Only applies to burned subtitles.

  Use "rgb", "rrggbb", or "aarrggbb" format, optionally prefixed with #. Three-digit colors are expanded to six digits. Colors without alpha are fully opaque. For eight-digit colors, the first two digits specify transparency ("00" is fully opaque, "ff" is fully transparent).

• font

  string

  The font family to use. Also includes boldness and style of the font. The default is Arial. With "style": "social", the default is Poppins, using its ExtraBold weight for JSON captions. An explicit font overrides the style.

  Here is a list of all supported fonts.

• font_color

  string | string(default: "FFFFFF")

  The color of the subtitle text. Only applies to burned subtitles.

  Use "rgb", "rrggbb", or "aarrggbb" format, optionally prefixed with #. Three-digit colors are expanded to six digits. Colors without alpha are fully opaque. For eight-digit colors, the first two digits specify transparency ("00" is fully opaque, "ff" is fully transparent).

• font_size

  string | number

  Specifies the size of the text in the 288-pixel subtitle coordinate space, scaled with the video height. The default is 16. With "style": "social", the default is 54. For social JSON captions, an omitted value scales this default with the shorter side of the video, increases it by 18% for portrait video, and shrinks captions as needed to fit at most two lines. Poppins capital letters are about 7.5% of the shorter side before fitting, or about 9% for portrait video. An explicit value keeps height-relative sizing and disables automatic fitting.

• bold

  boolean

  Specifies whether the subtitle text should be bold. Only applies to burned subtitles. The default is false. With "style": "social", the default is true; an explicit false overrides the style.

• italic

  boolean(default: false)

  Specifies whether the subtitle text should be italic. Only applies to burned subtitles.

• outline_width

  string | number | null

  Specifies the width of the text outline in the 288-pixel subtitle coordinate space, scaled with the video height. Numeric strings such as "4" are also accepted. Only applies to burned subtitles with "outline" or "box" border styles. If omitted or null, the default width is used (2 with "style": "social", proportional to the final fitted font size for social JSON captions).

• position

  bottom | bottom-left | bottom-right | center | left | right | top | 2 more ↓(default: "bottom")

  Specifies the position of the subtitles. With "style": "social", the default bottom position places JSON captions in the lower third, with extra bottom clearance on portrait video. An explicit value overrides the style.

• language

  string | null

  Specifies the language of the subtitle stream using a three-letter language code (e.g. "eng", "deu", "spa"). Only used if the subtitles are external. When adding multiple subtitle streams, set a different language for each stream so that video players can offer language selection.

• name

  string | null

  Specifies a human-readable name for the subtitle track (e.g. "English", "Deutsch", "Español"). Only used if the subtitles are external. This name is displayed in video player subtitle menus alongside the language.

• keep_subtitles

  boolean(default: false)

  Specifies if existing subtitles in the input file should be kept or be replaced by the new subtitle. Only used if the subtitles are external. Set this to true when chaining multiple /video/subtitle Steps to add several subtitle streams (e.g. different languages) to the same video.

Demos

Related blog posts