Understanding Language I/O credits

Language I/O uses a unified “credits” model to calculate how much to charge for translations requested through the various Language I/O applications and API. This means that, when you subscribe to a Language I/O plan, you get assigned a number of “credits” that you can then use for any version of flavor of translation. Language I/O then redeems these credits per translation, at a rate dependent on both the tier you subscribed to and the type of translation requested.

This system allows Language I/O customers to request translations of any type seamlessly, without the need to create a Purchase Order or quote.

Translation Types

There are seven types of translation. Translations are charged per word or per page, depending on the type. The types of translations are the following:

• Machine Translation (MT):
  This is the typical Machine Translation.
• Machine Translation of Attachments (MTA):
  Machine translation of attachments is charged per word or per page, depending on the file type of the attachment. For example:
  • Excel spreadsheets, for example, do not contain the concept of pages, and therefore the text is extracted, and charged per word at our standard MT rate.
  • On the other hand, PDF files have multiple pages, and therefore are charged per page.
  • Image files such as .jpg, .png, .gif are counted as a single page per file.
    For more information, see Translating attachments in Salesforce Case.
• Rapid Human Translation (RHT):
  This means speed-focused human translation services with a quick turnaround time (as opposed to traditional, quality-focused human translation that are more time-consuming).
• Machine Translation Post-Editing (MTPE):
  This means machine translations that are post-edited by a human linguist.
• MTPE+:
  Same as MTPE but for premium languages (see below).
• Human Translation (HT):
  This means traditional (quality-focused) human translation services (as opposed to rapid (speed-focused) human translation).
• HT+:
  Same as HT but for premium languages (see below).

Translation Memories

In addition, credit redemption is also affected by the use of translation memories:

• Translation Memory (TM):
  A TM is the database of the customer's already translated strings and segments. The translation memory is used by translators and automation tools in Language I/O to translate new strings and segments faster and consistently. Entries in the TM are called “segments”.
• Segment:
  A segment is the primary unit of translation. A segment can be a single word, short phrase, or a sentence. When content is sent to the Company for translation, the source content Segments are compared to Translation Memory to determine how much of the Segment and surrounding Segments have already been translated. This is called "Leverage". A higher Leverage can reduce the cost of the translation project.
• ICE Match:
  This means that an exact match has been found for a Segment in the Translation Memory and that this Segment occurs in exactly the same context (i.e., the same location) in a paragraph that is contained in Translation Memory. This context is defined by the surrounding segments. Some exceptions may apply.
• 100% Match:
  This means the text of a Segment passed to Language I/O for translation matches a Segment in the Translation Memory. It's considered a 100% match even if there are discrepancies with tags, placeholders, key values, etc. Some exceptions may apply.
• Fuzzy Match:
  This means that a segment has been translated previously in the same or a similar context as determined by Language I/O in its reasonable discretion, and the words therein will receive a discount on matches found in the Translation Memory as shown in the pricing table in your contract.

Premium languages

Premium languages are less common or rarer languages. They include: Norwegian, Swedish, Danish, Finnish, Japanese, Albanian, Catalan, Croatian, Hebrew, Icelandic, Indonesian, Maltese, Serbian, Slovenian, and Welsh.

Examples

Credits are redeemed per word or per page depending on the type of translation.

Example of rates and TM rates

Credits are redeemed differently depending on the translation type and language chosen. Here are three different scenarios that show how you could distribute 35,000 credits:

Choosing always Machine Translation (MT)

Choosing a mix of Machine Translation (MT) and Machine Translation Post-Editing (MTPE)

Choosing a mix of Machine Translation (MT) and Machine Translation of Attachments (MTA)

Different redemption rates also apply depending on your subscription tier.

Example for a chat conversation

• Messages between an agent and customer will be charged per word at the MT rate.
• If the end user sends a JPG as an attachment, the translation of the JPG is charged as MTA per page.

Example for an article translation

• Depending on the type of translation that you request, an article could be charged at either MT, MTPE or HT rates.
• Repetitions and fuzzy matches would apply at a discounted rate of the above per word rate, as described in the “Example of rates and TM rates” section.