One API, every medium.

1. 1.Upgrade to a plan that includes API access, then open Dashboard → API Keys and create a key.
2. 2.Send your first request with a Bearer token. Fast text translation returns the result inline.
3. 3.For documents and media, upload once and reference the file by id — or pass a hosted URL.
4. 4.Long-running jobs (dub, subtitle, large documents) return immediately as queued. Poll the job or receive a signed webhook when it completes.

Base URL — all endpoints are versioned under https://api.traxlate.com/v1

curl https://api.traxlate.com/v1/translations \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "target_lang": "es",
    "text": "The dashboard launches next week."
  }'

The resource-noun reference below is the full v1 design we're rolling out. The endpoints in this section are live right now under https://traxlate.com/api/v1 with the same Bearer tr_live_… auth. Copy-paste runnable.

Translate text

{ text, src, tgt } — short text returns inline; longer text returns a job id to poll.

curl https://traxlate.com/api/v1/translate \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "text": "The dashboard launches next week.", "src": "en", "tgt": "es" }'

Translate a batch of strings

{ texts: [], src, tgt } — for UI catalogs, CMS fields, and connectors. Returns { translations: [] } inline. Codes may be ISO (en) or FLORES (eng_Latn).

curl https://traxlate.com/api/v1/translate_batch \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "texts": ["Save", "Cancel", "Settings"], "src": "en", "tgt": "fr" }'

Translate a web page

{ url | html, src, tgt } — markup, links and shortcodes preserved; returns the translated HTML. Powers the WordPress plugin and the in-app website translator.

curl https://traxlate.com/api/v1/translate_html \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "url": "https://example.com/page", "src": "en", "tgt": "de" }'

Generate a presenter video

{ presenter, text, source_lang?, target_langs[], voice_id? } — renders a photoreal talking-head video, lip-synced, one per target language (the script is translated for each). Renders asynchronously; poll /v1/presenter/{id} and download the MP4 from /v1/presenter/{id}/video once status is done. Get the cast from /api/avatar/presenters.

# Create a talking-head video from text (one per target language)
curl https://traxlate.com/api/v1/presenter \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "presenter": "ava", "text": "Welcome to our product.", "source_lang": "en", "target_langs": ["es","fr"] }'

# Poll a job, then download when status == "done"
curl https://traxlate.com/api/v1/presenter/{id}    -H "Authorization: Bearer tr_live_..."
curl https://traxlate.com/api/v1/presenter/{id}/video -H "Authorization: Bearer tr_live_..." -o presenter.mp4

Order a human translation

{ text | file, source_lang, target_lang, tier, urgency } — a real professional linguist translates it. Priced per source word by tier (economy / professional / expert) and urgency (standard / express / urgent). Quote first, then place the order; credits are charged up front. Poll /v1/human/orders/{id} or subscribe to the human.delivered webhook to collect the result.

# Get an instant quote (priced per source word x tier x urgency)
curl https://traxlate.com/api/v1/human/quote \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{"text":"The full text to translate...","source_lang":"en","target_lang":"de","tier":"professional","urgency":"standard"}'
# -> { word_count, credits, usd_cents, promised_hours, ... }

# Place the order (charges credits up front; a professional linguist fulfils it)
curl https://traxlate.com/api/v1/human/orders \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{"text":"The full text...","source_lang":"en","target_lang":"de","tier":"professional","urgency":"express","instructions":"Formal tone; keep product names."}'
# Or upload a document (PDF/DOCX): multipart with a "file" field + the same params.

# Poll the order (or receive the human.delivered webhook)
curl https://traxlate.com/api/v1/human/orders/{id} -H "Authorization: Bearer tr_live_..."
# status: queued -> assigned -> in_progress -> delivered ; then delivered_text / delivered_file_ids

Automate it (n8n / Zapier / Make)

Every endpoint here is plain REST with a bearer token, and every order emits a signed webhook — so it drops straight into any workflow tool. A typical no-code flow: an HTTP Request node POSTs to /v1/human/orders (e.g. when a row is added to a sheet or a file lands in a folder), then a Webhook trigger listens for human.deliveredand writes the result back. Point your tool's webhook URL at the destination you register under Dashboard → API Keys → Webhooks; we sign every delivery with your key's secret.

Dub a video

Send the raw video as the request body; options ride in the query string: target_langs (plus optional source_lang, voice_mode, and lip_sync). Translates, clones the speaker's voice, and re-mixes — one job per target language. Set lip_sync=true to also re-animate the mouth to the new language (premium 2× tier). Renders asynchronously; poll /v1/dub/{id} and download the MP4 from /v1/dub/{id}/video once status is done.

# Dub a video into one or more languages.
# Send the raw video as the request body; options go in the query string.
# lip_sync=true re-animates the mouth to the new language (2x credits).
curl "https://traxlate.com/api/v1/dub?target_langs=es,ja&voice_mode=voice_clone&lip_sync=true" \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: video/mp4" \
  --data-binary @clip.mp4

# Poll a job, then download when status == "done"
curl https://traxlate.com/api/v1/dub/{id}        -H "Authorization: Bearer tr_live_..."
curl https://traxlate.com/api/v1/dub/{id}/video  -H "Authorization: Bearer tr_live_..." -o dubbed.mp4

Every request is authenticated with a Bearer token in the Authorization header. Live keys begin with tr_live_ and test keys with tr_test_. Keys are shown once on creation — store them securely; we cannot recover them.

Authorization: Bearer tr_live_<prefix><secret>

Test keys exercise the full API and return realistic responses without charging credits or producing billable renders. Rate limits are returned on every response via X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset. Exceeding the limit returns 429 with a Retry-After header.

There is one concept to learn. Translation, speech, dubbing, and subtitling all create a job — a resource with the same fields and the same lifecycle. Learn it once and every product works the same way.

A job moves through queued → processing → completed, or ends in failed / canceled. The id is prefixed by type (tr_, sp_, dub_, sub_), so you always know what you are holding.

{
  "id": "tr_3f9a21c8",
  "object": "translation",
  "status": "completed",
  "progress": 1.0,
  "created_at": "2026-05-19T10:00:00Z",
  "completed_at": "2026-05-19T10:00:02Z",
  "input": { "kind": "text", "detected_language": "en" },
  "output": {
    "target_lang": "es",
    "text": "El panel se lanza la próxima semana."
  },
  "error": null,
  "credits_charged": 5,
  "metadata": { "order_id": "1234" }
}

Fields

Synchronous vs. asynchronous

Fast work — short text translation and short speech — can return completed on the first response. Longer work — dubbing, subtitling, large documents — always returns 202 Accepted with a queued job and a Location header. You can force the async path on any request with "async": true.

HTTP/1.1 202 Accepted
Location: https://api.traxlate.com/v1/dubs/dub_a1b2c3

{
  "id": "dub_a1b2c3",
  "object": "dub",
  "status": "queued",
  "progress": 0,
  "created_at": "2026-05-19T10:00:00Z"
}

Retrieving, listing, and canceling

{resource} is one of translations, speech, dubs, subtitles. Poll a job every few seconds until it reaches a terminal state, or skip polling entirely with webhooks. Listing is cursor-paginated with ?limit= and ?starting_after=. Canceling a queued or in-progress job refunds reserved credits.

Idempotency

Send an Idempotency-Key header on any POST to make retries safe. A repeated key returns the original job instead of creating — and charging for — a duplicate.

Every product accepts input three ways. Pick whichever fits your integration:

You never declare a "mode." Send a file and the API detects its type and does the right thing — text is extracted from documents, written words are read from images, spoken words are recognized from audio and video. The job's input object always tells you what was detected.

Uploading a file

Upload once, reference anywhere. A single uploaded video can be both dubbed and subtitled without re-uploading. Large files upload in resumable parts. Files are retained for the job lifetime and then expire automatically.

curl https://api.traxlate.com/v1/files \
  -H "Authorization: Bearer tr_live_..." \
  -F "file=@quarterly_report.pdf" \
  -F "purpose=translation"

{
  "id": "file_8d2e",
  "object": "file",
  "filename": "quarterly_report.pdf",
  "kind": "pdf",
  "bytes": 184320,
  "expires_at": "2026-05-20T10:00:00Z"
}

Translate text, documents, or images into one or many languages. Pass an array of target languages to fan out — you get one batch with a child job per language, billed per target.

Body

Example — text, multiple targets

curl https://api.traxlate.com/v1/translations \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "source_lang": "auto",
    "target_lang": ["es", "fr", "ar"],
    "text": "The dashboard launches next week.",
    "glossary": ["dashboard = panel"]
  }'

Example — translate an uploaded document

curl https://api.traxlate.com/v1/translations \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "target_lang": "en",
    "file_id": "file_8d2e"
  }'

Document jobs add formatted exports to output (docx, pdf) alongside the translated text.

Multi-target response (batch)

{
  "id": "batch_77c0",
  "object": "batch",
  "status": "processing",
  "children": [
    { "id": "tr_a1", "target_lang": "es", "status": "completed" },
    { "id": "tr_a2", "target_lang": "fr", "status": "processing" },
    { "id": "tr_a3", "target_lang": "ar", "status": "queued" }
  ],
  "credits_charged": 15
}

Turn text into natural speech with a curated cast of named voices, optional voice cloning, emotion control, and a studio mode for the highest quality. Short renders return inline; longer scripts return a job to poll.

Body

Example

curl https://api.traxlate.com/v1/speech \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "voice": "aria",
    "target_lang": "en",
    "text": "Hello, world.",
    "mode": "studio"
  }'

Response

{
  "id": "sp_5b1d",
  "object": "speech",
  "status": "completed",
  "output": {
    "audio_url": "https://cdn.traxlate.com/...",
    "format": "wav",
    "duration_seconds": 1.4
  },
  "credits_charged": 50
}

Voice cloning

Upload a short reference clip to create a reusable cloned voice, then pass its clone_id to /v1/speech. Storing a clone is free; you pay only when you render with it. Cloned voices can speak any supported language.

Replace the spoken audio of a video or audio file with a translated voice track, timed to the original and matched to each speaker. Pass multiple target languages to produce several dubs from one upload. Always asynchronous.

Body

Example

curl https://api.traxlate.com/v1/dubs \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "file_id": "file_video01",
    "source_lang": "auto",
    "target_lang": ["es", "ja"],
    "voice_mode": "clone"
  }'

Completed output

{
  "id": "dub_a1b2c3",
  "object": "dub",
  "status": "completed",
  "input": { "kind": "video", "detected_language": "en", "duration_seconds": 612 },
  "output": {
    "target_lang": "es",
    "video_url": "https://cdn.traxlate.com/...mp4",
    "subtitles": {
      "srt": "https://cdn.traxlate.com/...srt",
      "vtt": "https://cdn.traxlate.com/...vtt"
    }
  },
  "credits_charged": 5100
}

Every completed dub also exposes matching subtitle files in output.subtitles at no extra cost — the dub already produced them.

Generate broadcast-ready captions from video or audio. Omit target_lang for a source-language transcript (the cheapest tier), or include one or more targets to also receive translated subtitles. Output is files only — no re-encoding, no burn-in.

Body

Example — caption + translate into FR & DE

curl https://api.traxlate.com/v1/subtitles \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "file_id": "file_video01",
    "source_lang": "auto",
    "target_lang": ["fr", "de"]
  }'

Completed output

{
  "id": "sub_9e0a",
  "object": "subtitle",
  "status": "completed",
  "output": {
    "target_lang": "fr",
    "srt": "https://cdn.traxlate.com/...fr.srt",
    "vtt": "https://cdn.traxlate.com/...fr.vtt",
    "txt": "https://cdn.traxlate.com/...fr.txt",
    "bilingual_srt": "https://cdn.traxlate.com/...bi.srt"
  },
  "credits_charged": 5400
}

Preview the exact credit cost of any job before you commit. No credits are charged. Send the same inputs you would send to the product endpoint, plus a product field.

curl https://api.traxlate.com/v1/quotes \
  -H "Authorization: Bearer tr_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "product": "dub",
    "file_id": "file_video01",
    "target_lang": ["es", "ja"]
  }'

{
  "product": "dub",
  "currency": "credits",
  "total": 10200,
  "lines": [
    { "target_lang": "es", "credits": 5100 },
    { "target_lang": "ja", "credits": 5100 }
  ]
}

/v1/languages lists supported languages per product with native names and tiers. /v1/voices lists the voice cast and the languages each can speak. /v1/me returns your account, current credit balance, and plan limits.

Skip polling. Register an endpoint in Dashboard → API Keys → Webhooks and we POST a signed event when a job changes state. Every event wraps the same job object you would get from a GET.

Payload

POST https://your-server.com/hooks/traxlate HTTP/1.1
Traxlate-Signature: t=1716112800,v1=<hmac-hex>
Content-Type: application/json

{
  "id": "evt_2a9c",
  "type": "job.completed",
  "created_at": "2026-05-19T10:00:09Z",
  "data": {
    "id": "dub_a1b2c3",
    "object": "dub",
    "status": "completed",
    "output": { "...": "..." },
    "credits_charged": 5100
  }
}

Events

Signature verification (Python)

import hmac, hashlib

def verify(payload: bytes, header: str, secret: str) -> bool:
    parts = dict(p.split("=", 1) for p in header.split(","))
    signed = f"{parts['t']}.".encode() + payload
    expected = hmac.new(secret.encode(), signed, hashlib.sha256).hexdigest()
    return hmac.compare_digest(expected, parts["v1"])

Always verify the Traxlate-Signature before processing. Respond 200 within 10 seconds; we retry failed deliveries with exponential back-off and keep delivery logs in the dashboard.

Errors share one envelope with a stable, machine-readable type and a request_id to quote in support requests.

{
  "error": {
    "type": "insufficient_credits",
    "message": "Your balance is too low to start this job.",
    "param": null,
    "request_id": "req_7f2a"
  }
}

The major version lives in the path (/v1). Additive changes — new fields, new event types — ship without a version bump, so write tolerant parsers. Breaking changes are introduced behind a dated Traxlate-Version header and announced in advance. Official Node and Python SDKs wrap this surface with typed helpers.