Translate text, HTML and JSON.Tags and keys come back intact.

Send a string, a web page or an i18n file to one REST API. AIbit translates the words and leaves markup, attributes, keys and protected values where they were. Name the engine per request, or let the API route it.

Median latency
146 ms
p95 290 ms
Languages
249
plus auto-detect
Success rate
99.98%
answered 200 OK
Requests
25.9M
in 7 days

Measured on 7 days of production traffic, 6-13 Sep 2026. Latency and success rate on text under 256 bytes.

Free tier with 1,000 requests a month, no credit card. Paid plans from $9.99/month.

Send text, HTML or JSON to the production API.

Choose any of 249 languages, name an engine or let the API route the request, and read the exact response body. Examples open with responses captured from production; Send runs them again, live, on a shared demo key.

POSThttps://edge.aibitranslator.com/api/v1/translator/html

engine omitted: /html runs on GTX, and a batch GTX fails is re-sent to GWEB.

207/500
12/12 tags200 OK
<div class="card">  <h3>Kabellose Ohrhörer</h3>  <p>    Bis zu <strong>30 Stunden</strong> der Akkulaufzeit.    <a href="/specs">Alle Spezifikationen ansehen</a>  </p>  <button data-sku="EB-220">In den Warenkorb legen</button></div>
Captured from production on 15 Sep 2026 / engine omitted

Examples for /html

Structure in, structure out.

Request and response bodies from the production API, captured on 15 Sep 2026 with the engine left out. Only words change: tags, attributes, keys and protected values come back where they were sent.

POST /json entoja

6/6 keys unchanged
request body
{"from": "en","to": "ja","json": {"cart": {"title": "Your cart","empty": "Your cart is empty.","checkout": "Proceed to checkout"},"brand": "Northwind","errors": ["Card declined","Address is required"]},"protected_paths": ["brand"]}
response body200 OK
{"trans": {"cart": {"title": "あなたのカート","empty": "カートは空です。","checkout": "チェックアウトに進む"},"brand": "Northwind","errors": ["カードが拒否されました","住所は必須です"]}}

Every key and the array stay in place. brand is listed in protected_paths, so it is never sent to an engine.

Run it in the playground

POST /html entofr

10/10 tags unchanged
request body, html field
<p>Hi <b>Maya</b>, your trial ends in <em>3 days</em>.</p><p><a href="https://app.example.com/billing">Choose a plan</a> to keep your projects.</p>
response body, trans field200 OK
<p>Salut <b>Maya</b>, votre essai se termine dans <em>3 jours</em>.</p><p><a href="https://app.example.com/billing">Choisissez un forfait</a> pour conserver vos projets.</p>

POST /text autotoen

request body
{"from": "auto","to": "en","text": "¿Dónde puedo cambiar la contraseña de mi cuenta?"}
response body200 OK
{"trans": "Where can I change my account password?","source_language_code": "es","source_language": "Spanish","trust_level": 1}

POST /text entovi

request body
{"from": "en","to": "vi","text": "Break a leg tonight!"}
response body200 OK
{"trans": "Chúc may mắn tối nay!"}

Name the engine per request, or let the API route it.

Every translate call takes an optional engine field. Leave it out for the routing below, or pin one engine when you need the same engine on every call.

A named engine runs the request or refuses it with 409. It is never swapped for another.

Refusals are answered before quota is counted, so a request the API will not serve costs nothing.

Engine values and the endpoints each one serves
engine/text/html/jsonWhat happens
omittedservedservedservedLeave engine out and the API routes the request: GCEV2 for /text, GTX then GWEB for documents.
"gcev2"served409, refused409, refusedThe /text default. Named on /html or /json, it is refused with 409.
"gtx"servedservedservedThe first engine for /html and /json: it carries the largest batch per upstream call.
"gweb"servedservedservedRe-translates a document batch GTX fails, or runs the whole request when named.

POST /text entode

honoured
request body
{"from": "en","to": "de","engine": "gtx","text": "Add to cart"}
response body200 OK
{"trans": "In den Warenkorb legen"}

POST /html entode

refused, not swapped
request body
{"from": "en","to": "de","engine": "gcev2","html": "<button>Add to cart</button>"}
response body409 Conflict
{"message": {"statusCode": 409,"message": "Engine 'GCEV2' does not serve the document endpoints (/json, /html); they run on GTX then GWEB. Name one of those, or omit 'engine' for the lane default.","error": "Conflict"},"statusCode": 409}

Pricing

Simple pricing that scales with you

Every plan includes full API access, all 249 languages and every translation engine. Start on the free tier, no setup fees, cancel anytime.

Pro

Perfect for side projects and small apps

$9.99/month

Side projects, MVPs, small SaaS apps

  • 100,000 requests / month
  • $0.00003 per extra request
  • 20 req/s rate limit
Get started
Recommended

MEGA

Best for growing startups and production apps

$120.00/month

Most popular — handles 150k daily users

  • 4,500,000 requests / month
  • $0.00003 per extra request
  • 200 req/s rate limit
Get started

Enterprise

High-volume apps with predictable costs

$200.00/month

Large-scale apps, no overage fees

  • 10,000,000 requests / month
  • No overage fees
  • 300 req/s rate limit
Get started

Unlimited

Dedicated server for mission-critical apps

$750.00/month

Enterprise apps, SLA guarantee, priority support

  • Unlimited requests
  • 500 req/s rate limit
Get started

Why developers choose AIbit

Compared with the most popular translation APIs

FeatureAIbit Translatorfrom $9.99/moGoogle Cloudpay-as-you-goDeepL APIfrom $5.49/mo
Cost per 1M chars$0.03$20.00$25.00
Free tier1,000 req/mo500k chars/mo500k chars/mo
JSON translationYesNoYes
HTML translationYesYesYes
Multiple translation engines4 enginesNoNo
No credit card to startYesNoNo

* Pricing based on public rates as of 2025. Google Cloud Translation standard rate $20/1M chars.

FAQ

Frequently asked questions

Have a question that is not here? The contact form at the bottom of the page reaches the team.

Why AIbit

Why AIbit Translator?

Built for developers who need reliable, fast, and affordable translation without vendor lock-in. Choose your preferred engine, pay only for what you use, and scale effortlessly.

249 Languages

Support for all major and minority languages worldwide. From English to Vietnamese, Japanese, French, Arabic, and beyond.

99.9% Uptime SLA

Enterprise-grade reliability with redundant infrastructure. Real-time status monitoring and automatic failover.

Developer-First Support

Fast email support, comprehensive API docs, and active community. Most queries answered within 24 hours.

Contact

Get in touch

Interested in a custom plan? Fill out the form below and we'll get back to you within 24 hours.