What you build<\/h2>\nThe playground experience can help you evaluate several aspects of translation behavior:<\/div>\n\n- How an LLM deployment translates a given language pair, including fluency characteristics<\/li>\n
- Whether a small number of reference translations may influence terminology, style, or tone<\/li>\n
- Whether an adaptive dataset can help guide domain context and support more consistent terminology usage across new text, depending on inputs<\/li>\n<\/ul>\n
Basic workflow<\/h3>\n\n- Open the Azure Translator Text translation model experience in Microsoft Foundry (Build > Models > AI Services > Azure Translator \u2013 Text translation).<\/li>\n
- Choose the source language, target language, and LLM deployment.<\/li>\n
- Translate a test sentence without adaptation and save the baseline result.<\/li>\n
- Add reference translation pairs or create an adaptive dataset from domain examples.<\/li>\n
- Translate the same sentence again and compare the output.<\/li>\n
- Apply the selected approach in your application using
referenceTextPairs<\/code> or adaptiveDatasetId<\/code> in the Translate API.<\/li>\n<\/ol>\nPrerequisites<\/h2>\nYou need:<\/div>\n\n- Microsoft Foundry project with access to Azure Translator Text translation<\/li>\n
- A Translator resource to programmatically use supported Text translation APIs<\/li>\n
- An LLM deployment available in the Translator model selection experience<\/li>\n
- A small set of high\u2011quality source and target sentence pairs for your domain<\/li>\n<\/ul>\nYou can use the languages API (with
scope=models<\/code>) to confirm available models programmatically.<\/div>\n<\/div>\n<\/div>\n\nimport requests\r\n\r\nurl = \"https:\/\/api.cognitive.microsofttranslator.com\/languages\"\r\nparams = {\r\n \"api-version\": \"2026-06-06\",\r\n \"scope\": \"models\",\r\n}\r\n\r\nresponse = requests.get(url, params=params)\r\nresponse.raise_for_status()\r\n\r\nprint(response.json())<\/code><\/pre>\n<\/div>\n\nPrepare the adaptive examples<\/h2>\nAdaptive translation is generally more effective when examples are clean, aligned, and similar to the text you plan to translate. In many cases, a small number of high\u2011quality examples can be more useful than a larger set of low\u2011quality examples.<\/div>\n<\/div>\nExample (English \u2192 Spanish):<\/strong><\/div>\n\n\n
- \n
- How an LLM deployment translates a given language pair, including fluency characteristics<\/li>\n
- Whether a small number of reference translations may influence terminology, style, or tone<\/li>\n
- Whether an adaptive dataset can help guide domain context and support more consistent terminology usage across new text, depending on inputs<\/li>\n<\/ul>\n
Basic workflow<\/h3>\n
- \n
- Open the Azure Translator Text translation model experience in Microsoft Foundry (Build > Models > AI Services > Azure Translator \u2013 Text translation).<\/li>\n
- Choose the source language, target language, and LLM deployment.<\/li>\n
- Translate a test sentence without adaptation and save the baseline result.<\/li>\n
- Add reference translation pairs or create an adaptive dataset from domain examples.<\/li>\n
- Translate the same sentence again and compare the output.<\/li>\n
- Apply the selected approach in your application using
referenceTextPairs<\/code> oradaptiveDatasetId<\/code> in the Translate API.<\/li>\n<\/ol>\nPrerequisites<\/h2>\n
You need:<\/div>\n- \n
- Microsoft Foundry project with access to Azure Translator Text translation<\/li>\n
- A Translator resource to programmatically use supported Text translation APIs<\/li>\n
- An LLM deployment available in the Translator model selection experience<\/li>\n
- A small set of high\u2011quality source and target sentence pairs for your domain<\/li>\n<\/ul>\nYou can use the languages API (with
scope=models<\/code>) to confirm available models programmatically.<\/div>\n<\/div>\n<\/div>\n\nimport requests\r\n\r\nurl = \"https:\/\/api.cognitive.microsofttranslator.com\/languages\"\r\nparams = {\r\n \"api-version\": \"2026-06-06\",\r\n \"scope\": \"models\",\r\n}\r\n\r\nresponse = requests.get(url, params=params)\r\nresponse.raise_for_status()\r\n\r\nprint(response.json())<\/code><\/pre>\n<\/div>\n\nPrepare the adaptive examples<\/h2>\n
Adaptive translation is generally more effective when examples are clean, aligned, and similar to the text you plan to translate. In many cases, a small number of high\u2011quality examples can be more useful than a larger set of low\u2011quality examples.<\/div>\n<\/div>\nExample (English \u2192 Spanish):<\/strong><\/div>\n\n\n
![\"Screenshot](\"https:\/\/devblogs.microsoft.com\/foundry\/wp-content\/uploads\/sites\/89\/2026\/06\/Screenshot-2026-05-14-at-9.46.00-PM-scaled.webp\")
<\/a><\/div>\n