Translation Memory is the single most valuable asset a professional translator builds over a career. Every approved translation is stored, reused, and made more valuable with each project. Here is a complete guide.
Translation Memory (TM) is a database that stores every approved source-target segment pair — when the same or similar text appears in a future project, the CAT tool retrieves the stored translation automatically, so translators never translate identical content twice and build an ever-growing asset of approved terminology.
Translation Memory (TM) is a database that stores source-language segments alongside their approved target-language translations. When the same or similar text appears in a future document, the CAT tool automatically retrieves and suggests the stored translation, eliminating the need to translate identical content again.
Translation Memory is the single most important efficiency tool in professional translation. It does not translate for you — it remembers what you have already translated and ensures you never have to do the same work twice. For a translator with years of work in a specific domain — pharmaceutical, legal, financial — the accumulated TM is effectively an institutional knowledge base of approved terminology and phrasing for that domain.
TM operates at the segment level. Most CAT tools split documents into segments at sentence boundaries. Each segment is stored as a source-target pair. When a new document contains a segment with 100% textual match to a stored pair, the translation is pre-filled automatically. When the match is partial (80–99%), the stored translation is suggested with the differences highlighted.
CAT tools classify TM matches into categories based on how closely the current segment matches a stored segment. The match percentage determines how much editing is required and typically influences the per-word rate charged for the work.
| Match Type | Score | What It Means | Editing Required | t09n.com Behaviour |
|---|---|---|---|---|
| 100% Match | 100% | Identical source segment found in TM | Verify only — no editing usually needed | Pre-filled automatically, does not count toward word balance |
| Context Match | 101% | Identical segment with identical context (surrounding segments also match) | Verify only | Pre-filled automatically, zero cost |
| Fuzzy Match | 75–99% | Very similar segment — minor differences in words, punctuation, numbers | Light editing of differing parts | Suggested with differences highlighted — counts as standard translation |
| Low Fuzzy | 50–74% | Partially similar segment — same structure, different content | Moderate editing or full re-translation | Suggested for reference only |
| No Match | 0–49% | No useful TM match found | Full translation required | MT output pre-filled (if MTPE enabled) or blank segment |
Translation Memory for Indian languages works the same way as for any other language pair — but the implementation quality varies significantly between tools. Generic CAT tools built for European languages may have issues with bidirectional text, native script character matching, and segment boundary detection for Indian language syntax.
t09n.com Translation Memory is built specifically for Indian language translation. Segments are stored in native script — Devanagari for Hindi and Marathi, Gujarati script for Gujarati, Tamil script for Tamil, Telugu script for Telugu. When a future document contains matching text, the retrieval works correctly regardless of script.
This matters practically for repetitive content types common in Indian language translation work:
Importing existing TM: If you have been working in memoQ or SDL Trados, you can export your Translation Memory as a TMX file and import it directly into t09n.com. Your accumulated translation work carries over — you do not start from zero.
These three terms get confused constantly because they all sit inside the same CAT tool workflow. They do different jobs:
In a typical t09n.com project: TM fills exact repeats for free, Glossary enforces terminology on every segment, and MTPE handles everything that's genuinely new. All three work together — none of them replaces the others.
TMX (Translation Memory eXchange) is the industry-standard XML file format for storing and exchanging Translation Memory between CAT tools. It's what makes TM portable — a TM built in memoQ or SDL Trados can be exported as a .tmx file and imported into a different tool without losing the source-target segment pairs.
A TMX file stores each segment pair with source language, target language, and the approved text — plus metadata like creation date and translator. Because it's a published, tool-agnostic standard, virtually every professional CAT tool supports TMX import and export, including t09n.com.
Translation Memory (TM) is a database that stores source segments alongside their approved target translations. When the same or similar text appears in a future document, the CAT tool automatically suggests the stored translation. It is the most important efficiency tool in professional translation — every translation you approve makes the TM more valuable for future projects.
TM reduces billable effort on repetitive content. In t09n.com, 100% TM matches are served free and do not count toward the word balance on any plan. For documents with high repetition — legal contracts, annual reports, product descriptions — TM leverage can reduce the chargeable word count by 30–70%, lowering cost per project significantly.
Yes. t09n.com TM fully supports Hindi, Gujarati, Tamil, Telugu, Kannada, Malayalam, Bengali, Marathi, Punjabi, Odia and Urdu. Segments are stored and retrieved in native script. The matching engine handles Indian language character encoding correctly.
Yes. You can export your TM from memoQ or SDL Trados as a TMX file and import it into t09n.com. Your accumulated translations carry over. TMX is the universal standard format for TM exchange across CAT tools.
TM leverage is the percentage of a document's word count covered by existing Translation Memory matches. A document with 70% TM leverage means 70% of segments have stored translations. High TM leverage reduces translation time and cost. For recurring clients, TM leverage typically grows with each project as more translations accumulate.
TMX (Translation Memory eXchange) is the industry-standard XML file format for storing and moving Translation Memory between CAT tools. Export your TM from memoQ or SDL Trados as a .tmx file and import it into t09n.com — the source-target segment pairs carry over intact.
Translation Memory reuses whole segments you've already approved. A Glossary enforces individual terms (brand names, product terminology) consistently across every segment, regardless of whether that segment came from TM or fresh translation. They solve different problems and are used together, not as alternatives.
t09n.com TM supports 10+ Indian languages. 100% matches always free. Start free — no credit card required.
Start Free →