Translator AI

Translator AI is an MCP server and CLI utility that automates the translation of JSON internationalization (i18n) files using artificial intelligence providers. It connects software projects to Google Gemini, OpenAI, or local Ollama instances such as DeepSeek-R1. Software developers, localization teams, and open-source maintainers use this tool to manage multilingual application dictionaries across single files or large nested directory trees. The tool parses JSON files, flattens paths, and intelligently batches translation requests while maintaining the original schema, arrays, nested objects, and template placeholders like URLs, emails, dates, and variables. Translator AI features incremental caching and multi-file deduplication, which prevent redundant API calls by translating only new or altered strings. It also includes utility options for alphabetical key sorting, automated source language detection, key presence verification to ensure no missing strings, and translation metadata tracking, enabling reliable and cost-effective continuous localization workflows across macOS, Windows, and Linux environments.

Category: Developer Tools & Code Intelligence

Tags: gemini, i18n, json, localization, ollama, translation

Visit Translator AI

How to install and configure Translator AI

Translator AI can be installed globally or locally via npm. Detailed instructions for configuring it as an MCP server with specific MCP client configurations are maintained in the project repository. 1. Install the package using npm: bash npm install -g translator-ai 2. Set up the required provider API key in your environment, such as GEMINI_API_KEY for Google Gemini or OPENAI_API_KEY for OpenAI, or prepare a local Ollama instance. 3. Refer to the repository documentation at https://github.com/DatanoiseTV/translator-ai for specific details on registering translator-ai within your chosen MCP client configuration file.

What you can do with Translator AI

• Translating software i18n JSON files across multiple target languages simultaneously using Google Gemini, OpenAI, or local Ollama models. - Preserving template placeholders, URLs, numbers, and nested JSON object hierarchies without manual string post-processing. - Running incremental translations during builds to localize only modified or recently added keys via built-in caching. - Deduplicating repeated strings across multiple locale source files to reduce translation API token consumption and costs. - Verifying translation integrity by checking that all source keys exist in the generated localized output files.

Key facts

• https://github.com/DatanoiseTV/translator-ai
• Developer Tools & Code Intelligence, Files, Documents & PDFs
• gemini, i18n, json, localization, ollama, translation

Part of MCP Servers

How do I install Translator AI?

You can install Translator AI via npm globally using npm install -g translator-ai, or locally in your project with npm install translator-ai. You will also need to configure your model provider credentials, such as GEMINI_API_KEY, OPENAI_API_KEY, or run a local Ollama instance.

What can Translator AI do?

Translator AI parses JSON i18n locale files and translates text into target languages using AI models. It preserves nested structure, arrays, and template variables, supports incremental caching to translate only altered strings, deduplicates keys across multiple files, and verifies that output files contain every expected key.

Which translation providers does Translator AI support?

Translator AI supports cloud-based models from Google Gemini and OpenAI, as well as local inference through Ollama using models like DeepSeek-R1. Provider selection and specific model names can be passed via command line flags or configured directly.

Does Translator AI alter non-translatable text?

When using format preservation options, Translator AI preserves URLs, email addresses, dates, numbers, and template variables without translating them, ensuring application placeholders remain functional across all localized output files.