Synthesize speech in documents
🤖/text/speak synthesizes speech in documents.

You can use the audio that we return in your application, or you can pass the audio down to other Robots to add a voice track to a video for example.
Another common use case is making your product accessible to people with a reading disability.
Transloadit aims to be deterministic, but this Robot uses third-party AI services. The providers (AWS, GCP, ElevenLabs) will evolve their models over time, giving different responses for the same input media. Avoid relying on exact responses in your tests and application.
Using ElevenLabs
Set provider to "elevenlabs" for ElevenLabs’ Multilingual v2 voices
(eleven_multilingual_v2), pinned so the output does not change under you. Every supported
language offers the female-1 and male-1 voices, each mapped to a voice that ElevenLabs
keeps available long-term. They speak the language of your text: ElevenLabs detects it, so
target_language only selects among the supported languages and cannot steer the pronunciation
of ambiguous words. ElevenLabs reads plain text only, so ssml must stay false, and it
speaks at most the first 10,000 characters.
For production, we recommend using your own ElevenLabs account. Store your key as
elevenlabs_api_key in a Template Credential of the AI Providers type and reference it with
credentials. With your own account, you have your own ElevenLabs limits and credits.
ElevenLabs bills you directly, and Transloadit charges only its fee: 30% of ElevenLabs’ list price
of $0.08 per 1,000 characters, at least $0.0005 per Step.
Transloadit’s shared ElevenLabs account is the default and a quick way to try things out. Its credits and capacity are shared by all customers, so your Step can be delayed or refused for reasons outside your control. With this account, Transloadit bills ElevenLabs’ list price of $0.08 per 1,000 characters plus a 30% fee, at least $0.05 per Step.
With either account, Transloadit retries HTTP 429 limit responses up to twice; if the limit persists, the Step fails. Exhausted credits fail without retrying. When the Step’s regular data volume charge is higher than the pricing above, that applies instead. If ElevenLabs rejects your key, the Step fails; Transloadit never falls back to its own key.
Supported languages and voices
AWS
| Language | Voices |
|---|---|
arb | female-1 |
cmn-CN | female-1 |
cy-GB | female-1 |
da-DK | female-1, male-1 |
de-DE | female-1, male-1 |
en-AU | female-1, male-1 |
en-GB | female-1, male-1 |
en-GB-WLS | male-1 |
en-IN | female-1 |
en-US | female-1, female-child-1, male-1, male-child-1 |
es-ES | female-1, male-1 |
es-MX | female-1 |
es-US | female-1, male-1 |
fr-CA | female-1 |
fr-FR | female-1, male-1 |
hi-IN | female-1 |
is-IS | female-1, male-1 |
it-IT | female-1, male-1 |
ja-JP | female-1, male-1 |
ko-KR | female-1 |
nb-NO | female-1 |
nl-NL | female-1, male-1 |
pl-PL | female-1, male-1 |
pt-BR | female-1, male-1 |
pt-PT | female-1, male-1 |
ro-RO | female-1 |
ru-RU | female-1, male-1 |
sv-SE | female-1 |
tr-TR | female-1 |
ELEVENLABS
| Language | Voices |
|---|---|
ar-XA | female-1, male-1 |
bg-BG | female-1, male-1 |
cmn-CN | female-1, male-1 |
cs-CZ | female-1, male-1 |
da-DK | female-1, male-1 |
de-DE | female-1, male-1 |
el-GR | female-1, male-1 |
en-AU | female-1, male-1 |
en-GB | female-1, male-1 |
en-US | female-1, male-1 |
es-ES | female-1, male-1 |
es-MX | female-1, male-1 |
fi-FI | female-1, male-1 |
fil-PH | female-1, male-1 |
fr-CA | female-1, male-1 |
fr-FR | female-1, male-1 |
hi-IN | female-1, male-1 |
hr-HR | female-1, male-1 |
id-ID | female-1, male-1 |
it-IT | female-1, male-1 |
ja-JP | female-1, male-1 |
ko-KR | female-1, male-1 |
ms-MY | female-1, male-1 |
nl-NL | female-1, male-1 |
pl-PL | female-1, male-1 |
pt-BR | female-1, male-1 |
pt-PT | female-1, male-1 |
ro-RO | female-1, male-1 |
ru-RU | female-1, male-1 |
sk-SK | female-1, male-1 |
sv-SE | female-1, male-1 |
ta-IN | female-1, male-1 |
tr-TR | female-1, male-1 |
uk-UA | female-1, male-1 |
GCP
| Language | Voices |
|---|---|
ar-XA | female-1, male-1 |
bn-IN | female-1, male-1 |
cmn-CN | female-1, male-1 |
cmn-TW | female-1, male-1 |
cs-CZ | female-1 |
da-DK | female-1, male-1 |
de-DE | female-1, male-1 |
el-GR | female-1 |
en-AU | female-1, male-1 |
en-GB | female-1, male-1 |
en-IN | female-1, male-1 |
en-US | female-1, female-2, female-3, male-1 |
es-ES | female-1, male-1 |
fi-FI | female-1 |
fil-PH | female-1, male-1 |
fr-CA | female-1, male-1 |
fr-FR | female-1, male-1 |
gu-IN | female-1, male-1 |
hi-IN | female-1, male-1 |
hu-HU | female-1 |
id-ID | female-1, male-1 |
it-IT | female-1, male-1 |
ja-JP | female-1, male-1 |
kn-IN | female-1, male-1 |
ko-KR | female-1, male-1 |
ml-IN | female-1, male-1 |
nb-NO | female-1, male-1 |
nl-NL | female-1, male-1 |
pl-PL | female-1, male-1 |
pt-BR | female-1 |
pt-PT | female-1, male-1 |
ro-RO | female-1 |
ru-RU | female-1, male-1 |
sk-SK | female-1 |
sv-SE | female-1 |
ta-IN | female-1, male-1 |
te-IN | female-1, male-1 |
th-TH | female-1 |
tr-TR | female-1, male-1 |
uk-UA | female-1 |
vi-VN | female-1, male-1 |
yue-HK | female-1, male-1 |
Usage example
Synthesize speech from uploaded text documents, using a female voice in American English:
{
"steps": {
"synthesized": {
"provider": "aws",
"robot": "/text/speak",
"target_language": "en-US",
"use": ":original",
"voice": "female-1"
}
}
}Parameters
interpolateboolean | 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
falseto treat every instruction field as literal text, or set an individual field path tofalseto treat only that field as literal text. For Robot-specific fields that are literal by default, set this totrueor set that field path totrueto opt back into interpolation.Use field names such as
path, or dotted paths such asffmpeg.vffor nested objects.output_metaRecord<string, boolean> | boolean | Array<string>Requests CPU-intensive metadata, disabled by default for faster Assemblies.
For images,
"has_transparency": truedetects transparency and"dominant_colors": trueextracts hex colors. Add"dominant_colors_sampled": truefor faster, approximate colors from downscaled frames. Sampling requiresdominant_colorsand changes neither image bytes nor billing. Its method and palette may vary. Omit or disable it for full-resolution colors.For images, you can also add
"blurhash": trueto 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": trueinstead extracts a base64-encoded ThumbHash intometa.thumbhash, together withmeta.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/handlefor uploaded originals or/image/resizefor processed outputs. Setting it only on/transloadit/storedoes 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": trueparameter to extract the colorspace of the output video.For videos, you can also add
"interlaced": trueto detect whether the video is interlaced. This combines the cheap ffprobefield_orderflag with a boundedidetsampling pass over the first frames of the source, exposinginterlaced,field_order, and a diagnosticinterlace_detectionobject underfile.meta. This is computationally expensive and billed accordingly.For audio, you can add
"mean_volume": trueto get a single value representing the mean average volume of the audio file.You can also set this to
falseto skip metadata extraction and speed up transcoding.user_metaRecord<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_metaon 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.resultboolean(default:false)Whether the results of this Step should be present in the Assembly Status JSON
queuebatchSetting 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_acceptboolean(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_acceptparameter set totrue, 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_errorsboolean | 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
trueis equivalent to["meta", "execute"]and will ignore errors in both phases.usestring | Array<string> | Array<object> | objectSpecifies 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
asto pass semantic intent to robots:{ "use": [ { "name": ":original", "as": "image" }, { "name": ":original", "as": "mask" } ] }
TipThat's likely all you need to know about
use, but you can view Advanced use cases.- You can pick any names for Steps except
promptstring | nullWhich text to speak. You can also set this to
nulland supply an input text file.providerauto | aws | elevenlabs | gcp(default:"auto")Chooses the best provider based on your request.
Select a specific provider to override automatic selection.
"auto"uses"aws"and never selects"elevenlabs".credentialsstringThe name of a Template Credential of the AI Providers type that holds your own
elevenlabs_api_key. Only the"elevenlabs"provider uses it.target_languagestring(default:"en-US")The written language of the document. This will also be the language of the spoken text.
The language should be specified in the BCP-47 format, such as
"en-GB","de-DE"or"fr-FR". Please consult the list of supported languages and voices.voicefemale-1 | female-2 | female-3 | female-child-1 | male-1 | male-child-1(default:"female-1")The gender to be used for voice synthesis. Please consult the list of supported languages and voices.
ssmlboolean(default:false)Supply Speech Synthesis Markup Language instead of raw text, in order to gain more control over how your text is voiced, including rests and pronounciations.
Please see the supported syntaxes for AWS and GCP. The
"elevenlabs"provider does not accept SSML.