MyMemory
App · 5 actions
Translate short texts, such as customer messages, support replies, product names and form answers, between more than 100 languages. MyMemory answers from its large shared memory of people's translations and falls back on machine translation. Translate one text, a list of texts, or one text into several languages at once, see every translation it holds for a sentence, and detect which language a text is in. No account needed.
Actions
| Action | What it does |
|---|---|
| Detect language | Find out which language a text is written in, with an English translation, for example to send each customer's message to the right team or to reply in their language. |
| Find translation matches | See every translation MyMemory holds for a sentence: people's translations of it and of similar sentences, and its machine translation, each with how closely it matches and who added it. |
| Translate a list of texts | Translate up to 50 texts into one language in one step, such as the answers to a form or a column of a spreadsheet, and get them back in the same order. |
| Translate into several languages | Translate one text into up to 10 languages in one step, such as a product name for each of your shops. |
| Translate text | Translate a text into another language, such as a customer's message, a support reply or a product name. |
Detect language
Find out which language a text is written in, with an English translation, for example to send each customer's message to the right team or to reply in their language. Only the opening of the text, about 500 letters, is sent. It needs a few words of real text: a lone word or a number can be misread. No account required.
Inputs
- Text
texttextRequired - The text whose language you want to know.
Outputs
characters_sentnumber- How many characters the step sent to MyMemory. They count toward its daily allowance: 5,000 characters a day without an email address on the connection, 50,000 with one.
english_translationtext- The checked text translated into English, or the text itself when it is already English.
is_englishtrue/false- True when the text is in English.
languageobject- The language the text is written in.
text_checkedtext- The part of the text that was sent: its opening, up to about 500 letters.
Example output
{
"characters_sent": 52,
"english_translation": "Where is my order? I've been waiting for two weeks.",
"is_english": false,
"language": {
"code": "de",
"name": "German"
},
"text_checked": "Wo ist meine Bestellung? Ich warte seit zwei Wochen."
}Find translation matches
See every translation MyMemory holds for a sentence: people's translations of it and of similar sentences, and its machine translation, each with how closely it matches and who added it. Use it to pick a wording yourself or to check how sure a translation is. One sentence of up to about 500 letters. No account required.
Inputs
- Text
texttextRequired - The sentence to look up, up to about 500 letters (fewer in scripts such as Chinese or Arabic).
- Translate from
from_languagetextRequired - The language the sentence is written in. MyMemory only searches its memory when it knows the language, so it can't be detected here.
- Translate into
to_languagetextRequired - The language to translate into. A language code from an earlier step, such as de or pt-BR, works too.
- Include machine translation
include_machine_translationtrue/false - Also return MyMemory's machine translation. Turn it off to see only translations people made. On by default.Default:
true.
Outputs
best_translationtext- The translation 'Translate text' would use: one that several people's translations of this exact sentence agree on, else the machine translation, else the closest match. Empty when there are none.
characters_sentnumber- How many characters the step sent to MyMemory. They count toward its daily allowance: 5,000 characters a day without an email address on the connection, 50,000 with one.
countnumber- How many translations were found. 0 is not an error.
from_languageobject- The language the sentence is in.
machine_translation_includedtrue/false- Whether MyMemory's machine translation was asked for.
matcheslist- Every translation MyMemory returned, closest first.
texttext- The sentence looked up.
to_languageobject- The language of the translations.
Example output
{
"best_translation": "Grazie per il tuo ordine",
"characters_sent": 24,
"count": 3,
"from_language": {
"code": "en",
"name": "English"
},
"machine_translation_included": true,
"matches": [
{
"added_by": "MateCat",
"created_at": "2023-05-08T23:17:55Z",
"from_locale": "en-GB",
"match_percent": 100,
"matched_text": "Thank you for your order",
"quality": 74,
"subject": null,
"times_used": 18,
"to_locale": "it-IT",
"translated_by": "human",
"translation": "Grazie per il tuo ordine",
"updated_at": "2023-05-08T23:17:55Z"
},
{
"added_by": "Public_Corpora",
"created_at": "2018-02-13T12:19:23Z",
"from_locale": "en-GB",
"match_percent": 100,
"matched_text": "Thank you for your order",
"quality": 0,
"subject": null,
"times_used": 2,
"to_locale": "it-IT",
"translated_by": "human",
"translation": "Grazie per l'ordine effettuato!",
"updated_at": "2018-02-13T12:19:23Z"
},
{
"added_by": "MateCat",
"created_at": "2023-05-09T19:13:24Z",
"from_locale": "en-GB",
"match_percent": 99,
"matched_text": "Thank you for your order.",
"quality": 74,
"subject": null,
"times_used": 2,
"to_locale": "it-IT",
"translated_by": "human",
"translation": "Grazie per il tuo ordine.",
"updated_at": "2023-05-09T19:13:24Z"
}
],
"text": "Thank you for your order",
"to_language": {
"code": "it",
"name": "Italian"
}
}Translate a list of texts
Translate up to 50 texts into one language in one step, such as the answers to a form or a column of a spreadsheet, and get them back in the same order. With Detect automatically, each text's language is worked out on its own, so answers written in different languages all come back in one. Blank items stay blank, and a text that appears twice is sent once. Up to 5,000 characters in all. No account required.
Inputs
- Texts
textslistRequired - The texts to translate, up to 50: a list from an earlier step, or one text per line.
- Translate from
from_languagetext - The language the text is written in. Leave it on Detect automatically and MyMemory works it out. A language code from an earlier step, such as fr or pt-BR, works too.Default:
auto. - Translate into
to_languagetextRequired - The language to translate into. A language code from an earlier step, such as de or pt-BR, works too.
Outputs
characters_sentnumber- How many characters the step sent to MyMemory. They count toward its daily allowance: 5,000 characters a day without an email address on the connection, 50,000 with one.
countnumber- How many texts there were, blank ones included.
to_languageobject- The language the texts were translated into.
translated_textslist- Just the translations, in the same order as the texts given: ready to write back to a spreadsheet column.
translationslist- Each text with its translation, in the order given.
Example output
{
"characters_sent": 56,
"count": 4,
"to_language": {
"code": "en",
"name": "English"
},
"translated_texts": [
"Very satisfied",
"Delivery was fast",
"",
"Very good service"
],
"translations": [
{
"already_in_target_language": true,
"from_language": {
"code": "en",
"name": "English"
},
"match_percent": null,
"original_text": "Very satisfied",
"translated_by": "unchanged",
"translated_text": "Very satisfied"
},
{
"already_in_target_language": false,
"from_language": {
"code": "fr",
"name": "French"
},
"match_percent": 85,
"original_text": "La livraison \u00e9tait rapide",
"translated_by": "machine",
"translated_text": "Delivery was fast"
},
{
"already_in_target_language": false,
"from_language": {
"code": null,
"name": null
},
"match_percent": null,
"original_text": "",
"translated_by": "unchanged",
"translated_text": ""
},
{
"already_in_target_language": false,
"from_language": {
"code": "es",
"name": "Spanish"
},
"match_percent": 85,
"original_text": "Muy buen servicio",
"translated_by": "machine",
"translated_text": "Very good service"
}
]
}Translate into several languages
Translate one text into up to 10 languages in one step, such as a product name for each of your shops. Each translation comes back in 'translations' and under its language code in 'by_language'. The text counts toward MyMemory's daily allowance once for each language. No account required.
Inputs
- Text
texttextRequired - The text to translate. It is sent once per language, so it can run to 5,000 characters divided by the number of languages.
- Translate from
from_languagetext - The language the text is written in. Leave it on Detect automatically and MyMemory works it out. A language code from an earlier step, such as fr or pt-BR, works too.Default:
auto. - Translate into
to_languageslistRequired - The languages to translate into, up to 10.
Outputs
by_languageobject- Each translation under its language code, e.g. 'fr', for picking one out in a later step.
characters_sentnumber- How many characters the step sent to MyMemory. They count toward its daily allowance: 5,000 characters a day without an email address on the connection, 50,000 with one.
countnumber- How many languages the text was translated into.
from_languageobject- The language the text was read as: the one picked, or the one MyMemory detected.
language_detectedtrue/false- True when MyMemory worked out the text's language.
original_texttext- The text that was translated, as given.
translationslist- One translation per language, in the order picked.
Example output
{
"by_language": {
"de": "Drahtlose Kopfh\u00f6rer",
"es": "Auriculares inal\u00e1mbricos",
"fr": "Ecouteur serre-t\u00eate sans fil",
"ja": "\u30ef\u30a4\u30e4\u30ec\u30b9\u30d8\u30c3\u30c9\u30d5\u30a9\u30f3"
},
"characters_sent": 76,
"count": 4,
"from_language": {
"code": "en",
"name": "English"
},
"language_detected": false,
"original_text": "Wireless headphones",
"translations": [
{
"already_in_target_language": false,
"language": {
"code": "fr",
"name": "French"
},
"match_percent": 100,
"translated_by": "human",
"translated_text": "Ecouteur serre-t\u00eate sans fil"
},
{
"already_in_target_language": false,
"language": {
"code": "de",
"name": "German"
},
"match_percent": 100,
"translated_by": "human",
"translated_text": "Drahtlose Kopfh\u00f6rer"
},
{
"already_in_target_language": false,
"language": {
"code": "es",
"name": "Spanish"
},
"match_percent": 100,
"translated_by": "human",
"translated_text": "Auriculares inal\u00e1mbricos"
},
{
"already_in_target_language": false,
"language": {
"code": "ja",
"name": "Japanese"
},
"match_percent": 85,
"translated_by": "machine",
"translated_text": "\u30ef\u30a4\u30e4\u30ec\u30b9\u30d8\u30c3\u30c9\u30d5\u30a9\u30f3"
}
]
}Translate text
Translate a text into another language, such as a customer's message, a support reply or a product name. Leave 'Translate from' on Detect automatically and MyMemory works out the language. You get the translation, whether a person or a machine made it, and how closely it matches. Up to 5,000 characters: MyMemory takes about 500 letters at a time, so longer text is split at its sentences, translated piece by piece and put back together. No account required.
Inputs
- Text
texttextRequired - The text to translate, up to 5,000 characters. Line breaks are kept.
- Translate from
from_languagetext - The language the text is written in. Leave it on Detect automatically and MyMemory works it out. A language code from an earlier step, such as fr or pt-BR, works too.Default:
auto. - Translate into
to_languagetextRequired - The language to translate into. A language code from an earlier step, such as de or pt-BR, works too.
Outputs
already_in_target_languagetrue/false- True when the text was already in the language to translate into, so it came back as it was.
characters_sentnumber- How many characters the step sent to MyMemory. They count toward its daily allowance: 5,000 characters a day without an email address on the connection, 50,000 with one.
from_languageobject- The language the text was read as: the one picked, or the one MyMemory detected. Empty when nothing needed translating.
language_detectedtrue/false- True when MyMemory worked out the text's language.
match_percentnumber- How closely the translation fits the text, from 0 to 100. 100 is a person's translation of this exact text; machine translations score 85; a lower score is a person's translation of a similar but different sentence. For a long text it is the lowest of its pieces. Empty when nothing needed translating.
original_texttext- The text that was translated, as given.
piecesnumber- How many pieces the text was translated in: 1 for a short text, more for a long one split at its sentences.
to_languageobject- The language the text was translated into.
translated_bytext- Who made the translation: 'human' (a person's translation from MyMemory's shared memory), 'machine' (machine translation), 'mixed' (a long text with some of each) or 'unchanged' (nothing needed translating).
translated_texttext- The translation.
Example output
{
"already_in_target_language": false,
"characters_sent": 68,
"from_language": {
"code": "fr",
"name": "French"
},
"language_detected": true,
"match_percent": 85,
"original_text": "Bonjour, ma commande n'est pas encore arriv\u00e9e. Pouvez-vous m'aider ?",
"pieces": 1,
"to_language": {
"code": "en",
"name": "English"
},
"translated_by": "machine",
"translated_text": "Hi, my order hasn't arrived yet. Can you help me?"
}