Medical and pharma translation is the part of the language industry where being wrong is expensive. A mistranslated dosage, a dropped exclusion criterion or an invented adverse-event term does not fail a style check, it fails a review, a submission or a patient. The good news is that most of the reference material that keeps that risk down is public and free, and a translator starting in the field can build a professional workflow without paying for a single licence.
This guide collects the free tools and sources that a medical or pharma translator actually uses: the terminology systems regulators enforce, the public termbases you can mine in a minute, the free QA and CAT options, and the clinical-trials sources where the source documents come from. It also says plainly where free stops: a dictionary is not a validation, and a licence gap is a commercial risk, not a terminology problem.
Every link below was opened and checked while writing this article, and the ones that are licensed rather than free are labelled as such. Nothing here replaces your own client glossary, the regulatory requirements of the target market, or the advice of a qualified regulatory professional.
In This Article
- What Makes Medical Translation Different
- Free Terminology Systems Regulators Use
- Free Termbases and Glossaries
- Free CAT and AI-Assisted Translation for Medical Work
- Free Quality Tools and Checklists
- Free Sources for Clinical Trials and Regulatory Documents
- A Free Medical Workflow, Step by Step
- Which Free Tool for Which Job
- What Free Tools Do Not Do
- Free Courses and Training at TranslaStars
- Frequently Asked Questions
1. What Makes Medical Translation Different
Three things separate this specialisation from general translation, and all three shape the toolkit.
The terminology is standardised, and the standardisation is public. Drug names follow the WHO International Nonproprietary Names, adverse events follow the MedDRA hierarchy, procedures and diagnoses follow ICD and SNOMED CT, laboratory observations follow LOINC. A translator who works from the same systems the sponsor uses stops arguing about vocabulary and starts checking it.
The register is regulated. Patient information leaflets, summaries of product characteristics, informed consent forms and labels have their own conventions in every market, and the wording of a patient-facing text is often constrained by readability rules. Free public documents, especially the published product information from the EMA and the FDA, are the best style guide available because they are the same genre you are asked to produce.
The risk is asymmetric. In marketing translation a clumsy sentence costs a client; in a clinical trial document the same sentence costs a protocol deviation. That is why the free checklist culture of this field (peer review, back translation, reconciliation, linguistic validation) matters more than any single tool.
Free tools lower the cost of the research. They do not lower the standard of the deliverable, and no public database releases you from your client glossary or the market regulator.
AI-Assisted Medical Translation with Wordscope
The free CAT tool of this article, applied to medical documents, with the anonymisation and terminology settings that matter.
Free2. Free Terminology Systems Regulators Use
Start with the systems, not with a dictionary. These are the reference vocabularies that appear in the documents themselves, and most of them are free to search online.
2.1. WHO International Nonproprietary Names (INN)
The INN system gives every active substance a single global name, which is why the same molecule is not a different drug in every market. Use it to decide the stem of a drug name (-mab, -tinib, -prazole) and to check that a brand name in your source is not being used where the substance is meant. The lists and the naming rules are public.
2.2. MeSH and the UMLS
MeSH is the National Library of Medicine's controlled vocabulary for indexing biomedical literature; it is a fast way to find the accepted English term for a concept and its synonyms. The UMLS Metathesaurus goes further and maps many of these vocabularies to each other, including multilingual sources. Both are free to consult; the UMLS download requires a licence and an account, and that licence is what also unlocks SNOMED CT in the United States.
2.3. SNOMED CT, LOINC and ICD-11
SNOMED CT is the clinical terminology behind electronic health records in a large number of countries, with national translations that are often the best source for a clinical term nobody has written down in a glossary yet. In the United States it is distributed to holders of the NLM UMLS licence, and national release centres distribute it under their own terms; check the terms that apply in your country before you redistribute anything. LOINC is the equivalent for laboratory observations, and it is the tool to reach for when a lab test name looks improvised. The ICD-11 browser from the WHO gives classifications, coding guidance and multilingual views in one place.
2.4. MedDRA: the one that is licensed
MedDRA is the medical dictionary for regulatory activities: adverse events, indications, investigations and the whole hierarchy that pharmacovigilance runs on. It is not a free-for-all resource. Access is granted through a subscription managed by the MSSO, with different terms for commercial, academic, regulatory and non-profit users, so check the MedDRA FAQ and the subscription terms before you use its terms in a deliverable. Knowing the system exists, and knowing which version a client is on, is free; using it is a licence question.
2.5. Patient-facing and professional references
For the plain-language end of the work, the NCI Dictionary of Cancer Terms, MedlinePlus and the MSD Manuals Professional edition are all public, and all three are written in the register your patient-facing target text needs. When a sentence has to be understandable to a patient with no medical training, reading how these sources phrase the same concept is more useful than reading another translation.
Pharmacovigilance adds one more free door: the Uppsala Monitoring Centre is the WHO centre for international drug monitoring and publishes the reference material behind the WHODrug dictionaries and the causality assessment vocabulary used in case narratives.
3. Free Termbases and Glossaries
Systems tell you what a term means. Termbases tell you what a term is called in two languages, which is the part that saves the hour.
3.1. IATE and UNTERM
IATE is the European Union's interinstitutional terminology database: searchable, multilingual, and strongest exactly where public-health and pharmaceutical regulation is written at EU level. UNTERM does the same job for the United Nations system and is useful for global health language, including the terminology used by WHO documents.
3.2. Termium Plus
The Government of Canada's Termium Plus is free, bilingual (English, French, Spanish and Portuguese records) and unusually good on regulated language, because it records the writing conventions of Canadian health and drug regulation. When your target is French or Canadian French, it is often the only free source that resolves a labelling term cleanly.
3.3. OneClickTerms
When the client sends you a pile of parallel documents and no glossary, OneClickTerms extracts candidate bilingual terminology from a bilingual corpus for free. It is the fastest way to build a first-pass termbase from a set of aligned PDFs, and its output is a proposal to review, not a glossary to trust blindly.
3.4. My Med Pharm
My Med Pharm is a catalogue of more than 500 medical and pharmaceutical resources, put together by Ekaterina Chashnikova. It gathers the sites, tools, glossaries, corpora and podcasts that medical translators already use, with notes on how to get the most out of each one, and it is open for translators to add their own finds. It also carries a blog on the terminology questions that come up in real medical and pharma work, which is often the fastest way back to a documented source.
4. Free CAT and AI-Assisted Translation for Medical Work
You do not need a paid CAT licence to run a medical project, but you do need one that handles terminology, tags and a format the client accepts.
4.1. Wordscope
Wordscope is a browser-based professional CAT tool with AI and machine translation assistance, translation memories, terminology databases, quality tools, client-specific customisation and anonymisation, and it is the tool our own medical course is built around. There is nothing to install, a project can be created without a credit card, and for a medical translator the anonymisation features matter more than the automation.
OmegaT is the long-standing open-source option for translators who want a local tool with their own TM on disk, and we compare it with the other free tools in Best Free CAT Tools in 2026. MateCat, Smartcat and Crowdin all offer free access with cloud projects, which is often the quickest way to share a medical glossary with a reviewer in another country without sending files around.
For document conversion and format work, the Okapi Framework is free and open source, and it is the tool to use when a sponsor sends an InDesign file, an XLIFF with unusual segmentation or a batch of DOCX files that has to become one clean bilingual table. It also carries the CheckMate QA module, which we come back to below.
Machine translation is a drafting aid in this field, never an authority. In a pharmacovigilance or clinical document, an MT engine that renders an adverse-event term with a synonym instead of the controlled term creates the exact error the reviewer is looking for.
5. Free Quality Tools and Checklists
Xbench is the free quality and terminology tool that most medical translators end up keeping installed: it loads TMX, XLIFF, Trados, Wordfast, memoQ and other bilingual formats, searches all of them with one hotkey, spell-checks hundreds of files at once, and runs predefined checks for completeness, consistency, numbers, tags and key terms. The download page publishes the free development builds, and the checklists are the free part that matters: writing your own list of medical checks (units, dosage strings, numbers, contraindication wording, product names) turns a general QA tool into a medical one.
Okapi CheckMate runs a comparable battery of checks inside the free Okapi Framework and is easier to script for batch work. Manual checklists remain the strongest free instrument in this field, and the ones we teach in our life-sciences courses cover five groups: numbers and units, terminology against the client glossary, tag and placeholder integrity, consistency across the document set, and the register rules of the target market.
| Check | What it catches | Free tool |
|---|---|---|
| Numbers and units | Dosage, concentrations, dates, ranges and decimal separators that changed in translation | Xbench, Okapi CheckMate, a manual pass |
| Terminology consistency | The same concept translated three ways across a protocol, an IB and a consent form | Xbench key-term check, Wordscope termbases |
| Tag and placeholder integrity | Lost variables in a template, broken links, tags moved inside a sentence | Xbench, Okapi CheckMate |
| Untranslated and missing segments | Empty targets, copied source, segments that never came back from the reviewer | Xbench completeness check |
| Register and readability | Patient-facing text that reads like a scientific abstract | Manual review, public patient information from EMA and FDA |
6. Free Sources for Clinical Trials and Regulatory Documents
The documents you translate are usually public before they reach you, which makes research a free and fast step rather than a guess.
ClinicalTrials.gov holds registrations, protocols, statistical analysis plans in some cases, and results summaries for trials worldwide. For trials run in the European Union, the Clinical Trials Information System (CTIS) is the public register and the source of the harmonised EU documentation set. Between the two of them you can often read the same protocol in two languages before you translate it.
For the product side, the European Medicines Agency publishes the European public assessment reports and product information that define the genre, the Heads of Medicines Agencies network covers the national procedures in the same family, and in the United States Drugs@FDA gives the approved labelling and review documents. The ICH guidelines, free to read, define the structure of the documents you will be translating (the common technical document, safety reporting, good clinical practice).
Two more free references are worth bookmarks. ISPOR publishes the good-practice material behind patient-reported outcome translation and linguistic validation, which is the workflow a lot of clinical-trial translators end up inside. EFSA covers the food-safety side of the same industry, where claims and dossiers have their own vocabulary.
Linguistic Validation in Clinical Trials and Practice
The validation method behind patient-reported outcomes, consent forms and instruments used across markets.
Free7. A Free Medical Workflow, Step by Step
Here is the workflow I would hand to a translator starting with medical clients tomorrow, using only the free options above.
- Read the genre before the text. Open the published product information for the same kind of product in your target language (EMA or Drugs@FDA) and note the fixed phrases of the genre.
- Build the client glossary first. Pull the terms from the client, then fill the gaps from WHO INN, MeSH, ICD-11 and the national translation of SNOMED CT where it exists. Record the source of every entry, because a reviewer will ask.
- Extract what the client did not give you. Run parallel documents through OneClickTerms and check the candidates against IATE, UNTERM or Termium Plus.
- Translate in a tool that holds the terminology. Wordscope or any free CAT tool, with the termbase loadable and the source and target side by side.
- Convert and clean the format with Okapi if the client sent files a CAT tool cannot ingest directly.
- Run the mechanical QA in Xbench (numbers, tags, untranslated, key terms) before you read the text as a reader.
- Do the manual pass out loud on the patient-facing parts, then hand the document to a second pair of eyes when the client allows it.
8. Which Free Tool for Which Job
| Job | Free option | When to move up |
|---|---|---|
| Find the accepted English term | MeSH, UMLS, WHO INN | When the client enforces a controlled dictionary such as MedDRA |
| Check a bilingual term | IATE, UNTERM, Termium Plus | When the client has its own termbase and it wins |
| Extract terminology from parallel files | OneClickTerms | When the corpus is large enough to need a professional terminology workflow |
| Translate with terminology support | Wordscope, OmegaT, MateCat, Smartcat, Crowdin | When the client requires a specific TMS or a connector |
| Convert and clean file formats | Okapi Framework | When the volume of desktop publishing work justifies a paid conversion suite |
| Quality check a delivery | Xbench, Okapi CheckMate, manual checklists | When the client asks for a signed QA report or a validated workflow |
| Read the source genre | EMA, Drugs@FDA, ClinicalTrials.gov, CTIS, ICH | Never: this part stays free |
9. What Free Tools Do Not Do
Four limits worth stating clearly, because each one is a real project risk.
A public dictionary is not a validation. SNOMED CT, LOINC and ICD give you terminology, not a marketing authorisation. Nothing in this article says that a term you found is the term a regulator will accept for a specific label.
Licences are not uniformity. MedDRA and some national terminology releases are licensed, and the terms differ between commercial, academic and regulatory users. Using a licensed dictionary without a subscription is a commercial risk, not a shortcut.
Language coverage is uneven. Some terminology systems are fully translated for a handful of languages and barely represented in others. Where no official translation exists, the honest deliverable is the term plus a documented note, not an invented translation.
Free tools do not review your judgement. The QA tools catch inconsistency and numbers. They cannot tell you that a sentence in a consent form has become a promise, or that a contraindication reads weaker than the source. That is the specialist's fee, and it is not the part a free tool replaces.
10. Free Courses and Training at TranslaStars
Three of our life-sciences courses cost nothing, and they pair with this toolkit exactly: one gives you the projects, one the clinical-trial method, one the AI-assisted tooling.
Life Sciences Translation: Projects & Expertise
The projects a life-sciences translator is actually asked to deliver, and the expertise each one demands.
FreeThe other two courses are placed beside the sections they support, and the full catalogue is at TranslaStars courses.
When you want the structured route, these are the paid programmes that pick up where the free material ends: Fundamentals of Medicine for Translators, Pharmacology for Translators and Interpreters (II Ed.), Clinical Trials Explained for Translators (Ed. II), Introduction to Medical Device Localization, Medical Reports: Translation & Editing Mastery and Marketing Translation for Pharmaceutical Companies (Ed. II). For the full path, including project management and AI pipelines for this sector, see the Master in AI, Translation & PM for Life Sciences. The complete catalogue is on the TranslaStars courses page, and the free ones are collected with the rest of our free material at free translation courses and resources.
Frequently Asked Questions
10.1. Can I work in medical translation with only free tools?
For research, terminology, translation, conversion and mechanical QA, yes: the whole chain in this article is free. What is not free is expertise, and the point where money changes hands is usually the controlled dictionary (MedDRA in particular), a client-mandated TMS, or the training that teaches you to read a protocol. The free tools remove the excuse, not the requirement.
10.2. Is MedDRA free?
Not for general use. Access goes through a subscription administered by the MSSO, with different terms for commercial, academic, regulatory and non-profit users. Search the MedDRA site for the current terms rather than assuming a category covers you.
10.3. Which free source is best for patient-facing texts?
Published patient information leaflets and public patient-facing material from EMA and the FDA, plus plain-language references such as MedlinePlus and the NCI Dictionary of Cancer Terms. They are written in the register your target text needs, which no terminology database will teach you.
10.4. Do free translation memories and glossaries break confidentiality?
Cloud tools store what you upload, and some of them use client content according to their own terms. Before you put a trial document, a patient leaflet in draft or anything under an NDA into a free cloud tool, check the data-processing terms and the client's own security requirements. Wordscope's anonymisation features exist for this reason, and a local tool such as OmegaT or Okapi keeps the content on your machine.
10.5. Is machine translation safe for medical documents?
As a drafting aid, it saves time; as a deliverable, it is a risk unless a qualified translator has checked every number, unit, term and negated statement. Adverse-event terminology, dosage strings and negation are exactly what MT gets wrong in a way that matters.
10.6. Where do I start if I have no medical clients yet?
Pick one genre (product information is the friendliest), read twenty published examples in your target language, build a 200-term bilingual glossary from the free sources above, and translate three sample documents to professional standard. That body of work is what a life-sciences client asks to see before a paid test.
More on this blog: How to Break into Life Sciences Localization for the career path, Best Free CAT Tools in 2026 for the translation environment, and 10 Best AI Tools for Language Professionals for the wider AI toolkit.
