BCP-47
Also written: RFC 5646
A standard for identifying languages, using tags to specify language, region, or dialect.
BCP-47 is a set of rules for creating codes that identify languages. Think of it like a shorthand for saying which language a piece of text or speech is in.
For example, 'en-US' means English spoken in the United States, while 'es-ES' means Spanish spoken in Spain. This helps computers know how to process and display text correctly, especially when dealing with multiple languages.
BCP-47 (Best Current Practice 47) is the IETF standard for language identification. It defines a syntax for language tags, which are composed of one or more 'subtags' separated by hyphens. Primary subtags identify a language (e.g., 'en' for English), while optional subtags can specify a script (e.g., 'Latn' for Latin script), region (e.g., 'US' for United States), or variant (e.g., 'macos' for a specific macOS variant).
This standard is crucial for internationalization and localization, enabling systems to correctly select fonts, sort text, and apply language-specific rules. It is widely used in web standards (HTML, XML), operating systems, and software applications, including AI models that process and generate multilingual content.
BCP-47, also known as RFC 5646, is a standard for language identification. Frontier AI labs like Google Gemini are using it to specify the languages their models support, especially in applications like speech-to-text where accurate language detection is crucial for performance and user experience.
- speech_to_text Gemini 3.5 TranscribeGoogle · Gemini Models ·
- Binjakët 3.5 TranskriptimGoogle · Gemini Models ·
- Google Gemini
- Google Gemini uses BCP-47 tags to identify and support a wide range of languages in its speech-to-text services, ensuring accurate transcription across different linguistic variations.
- ISO 639
- A standard that defines language codes, often used as primary subtags w
- ISO 3166
- A standard that defines country codes, often used as region subtags wi
- RFC 5646
- BCP-47 is defined by RFC 5646, which specifies the format for language
BCP-47 is the Internet Engineering Task Force's (IETF) recommended practice for language identification. It is defined in RFC 5646, which updates and obsoletes previous standards like RFC 4646. Source ↗
When using a multilingual AI tool, try inputting text or speech in a language other than your primary one. Observe if the AI correctly identifies the language and responds appropriately. For example, ask Google Gemini to transcribe a sentence in Spanish and then one in French.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- What is the difference between BCP-47 and ISO codes?
- BCP-47 is a framework for language tags that often uses ISO 639 (language) and ISO 3166 (region) codes as components. BCP-47 tags can include more than just language and region, like scripts or variants.
- Why is language identification important for AI?
- Accurate language identification helps AI models process text and speech correctly, leading to better performance in tasks like translation, transcription, and content generation across different languages.
First seen on Subvolts Sep 12, 2026. Updated Sep 30, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.