Lara Translate MCP Server

A Model Context Protocol (MCP) Server for Lara Translate API, enabling powerful translation capabilities with support for language detection, context-aware translations and translation memories.

📚 Table of Contents

• 📖 Introduction
• 🛠 Available Tools
• 🚀 Getting Started
  • 📋 HTTP Server
  • 🔌 STDIO Server
• 🧪 Verify Installation
• 💻 Popular Clients that supports MCPs
• 🆘 Support

📖 Introduction

<details> <summary><strong>What is MCP?</strong></summary>

Model Context Protocol (MCP) is an open standardized communication protocol that enables AI applications to connect with external tools, data sources, and services. Think of MCP like a USB-C port for AI applications - just as USB-C provides a standardized way to connect devices to various peripherals, MCP provides a standardized way to connect AI models to different data sources and tools.

Lara Translate MCP Server enables AI applications to access Lara Translate's powerful translation capabilities through this standardized protocol.

More info about Model Context Protocol on: https://modelcontextprotocol.io/

</details> <details> <summary><strong>How Lara Translate MCP Works</strong></summary>

Lara Translate MCP Server implements the Model Context Protocol to provide seamless translation capabilities to AI applications. The integration follows this flow:

1. Connection Establishment: When an MCP-compatible AI application starts, it connects to configured MCP servers, including the Lara Translate MCP Server
2. Tool & Resource Discovery: The AI application discovers available translation tools and resources provided by the Lara Translate MCP Server
3. Request Processing: When translation needs are identified:
   • The AI application formats a structured request with text to translate, language pairs, and optional context
   • The MCP server validates the request and transforms it into Lara Translate API calls
   • The request is securely sent to Lara Translate's API using your credentials
4. Translation & Response: Lara Translate processes the translation using advanced AI models
5. Result Integration: The translation results are returned to the AI application, which can then incorporate them into its response

This integration architecture allows AI applications to access professional-grade translations without implementing the API directly, while maintaining the security of your API credentials and offering flexibility to adjust translation parameters through natural language instructions.

</details> <details> <summary><strong>Why to use Lara inside an LLM</strong></summary>

Integrating Lara with LLMs creates a powerful synergy that significantly enhances translation quality for non-English languages.

Why General LLMs Fall Short in Translation

While large language models possess broad linguistic capabilities, they often lack the specialized expertise and up-to-date terminology required for accurate translations in specific domains and languages.

Lara’s Domain-Specific Advantage

Lara overcomes this limitation by leveraging Translation Language Models (T-LMs) trained on billions of professionally translated segments. These models provide domain-specific machine translation that captures cultural nuances and industry terminology that generic LLMs may miss. The result: translations that are contextually accurate and sound natural to native speakers.

Designed for Non-English Strength

Lara has a strong focus on non-English languages, addressing the performance gap found in models such as GPT-4. The dominance of English in datasets such as Common Crawl and Wikipedia results in lower quality output in other languages. Lara helps close this gap by providing higher quality understanding, generation, and restructuring in a multilingual context.

Faster, Smarter Multilingual Performance

By offloading complex translation tasks to specialized T-LMs, Lara reduces computational overhead and minimizes latency—a common issue for LLMs handling non-English input. Its architecture processes translations in parallel with the LLM, enabling for real-time, high-quality output without compromising speed or efficiency.

Cost-Efficient Translation at Scale

Lara also lowers the cost of using models like GPT-4 in non-English workflows. Since tokenization (and pricing) is optimized for English, using Lara allows translation to take place before hitting the LLM, meaning that only the translated English content is processed. This improves cost efficiency and supports competitive scalability for global enterprises.

</details>

🛠 Available Tools

Translation Tools

<details> <summary><strong>translate</strong> - Translate text between languages</summary>

Inputs:

• text (array): An array of text blocks to translate, each with:
  • text (string): The text content
  • translatable (boolean): Whether this block should be translated
• source (optional string): Source language code (e.g., 'en-EN')
• target (string): Target language code (e.g., 'it-IT')
• context (optional string): Additional context to improve translation quality
• instructions (optional string[]): Instructions to adjust translation behavior
• source_hint (optional string): Guidance for language detection
• glossaries (optional string[]): Array of glossary IDs to enforce terminology (e.g., ['gls_xyz123'])
• no_trace (optional boolean): Privacy flag - if true, request won't be traced/logged
• priority (optional string): Translation priority - 'normal' or 'background'
• timeout_in_millis (optional number): Custom timeout in milliseconds

Returns: Translated text blocks maintaining the original structure

</details>

Glossaries Tools

<details> <summary><strong>list_glossaries</strong> - List all glossaries</summary>

Inputs: None

Returns: Array of glossaries with their details (id, name, createdAt, updatedAt, ownerId)

</details> <details> <summary><strong>get_glossary</strong> - Get a specific glossary by ID</summary>

Inputs:

• id (string): The glossary ID (e.g., 'gls_xyz123')

Returns: Glossary object or null if not found

</details>

Translation Memories Tools

<details> <summary><strong>list_memories</strong> - List saved translation memories</summary>

Returns: Array of memories and their details

</details> <details> <summary><strong>create_memory</strong> - Create a new translation memory</summary>

Inputs:

• name (string): Name of the new memory
• external_id (optional string): ID of the memory to import from MyMemory (e.g., 'ext_my_[MyMemory ID]')

Returns: Created memory data

</details> <details> <summary><strong>update_memory</strong> - Update translation memory name</summary>

Inputs:

• id (string): ID of the memory to update
• name (string): The new name for the memory

Returns: Updated memory data

</details> <details> <summary><strong>delete_memory</strong> - Delete a translation memory</summary>

Inputs:

• id (string): ID of the memory to delete

Returns: Deleted memory data

</details> <details> <summary><strong>add_translation</strong> - Add a translation unit to memory</summary>

Inputs:

• id (string | string[]): ID or IDs of memories where to add the translation unit
• source (string): Source language code
• target (string): Target language code
• sentence (string): The source sentence
• translation (string): The translated sentence
• tuid (optional string): Translation Unit unique identifier
• sentence_before (optional string): Context sentence before
• sentence_after (optional string): Context sentence after

Returns: Added translation details

</details> <details> <summary><strong>delete_translation</strong> - Delete a translation unit from memory</summary>

Inputs:

• id (string): ID of the memory
• source (string): Source language code
• target (string): Target language code
• sentence (string): The source sentence
• translation (string): The translated sentence
• tuid (optional string): Translation Unit unique identifier
• sentence_before (optional string): Context sentence before
• sentence_after (optional string): Context sentence after

Returns: Removed translation details

</details> <details> <summary><strong>import_tmx</strong> - Import a TMX file into a memory</summary>

Inputs:

• id (string): ID of the memory to update
• tmx_content (string): The content of the tmx file to upload
• gzip (boolean): Indicates if the file is compressed (.gz)

Returns: Import details

</details> <details> <summary><strong>check_import_status</strong> - Checks the status of a TMX file import</summary>

Inputs:

• id (string): The ID of the import job

Returns: Import details

</details>

🚀 Getting Started

Lara supports both the STDIO and streamable HTTP protocols. For a hassle-free setup, we recommend using the HTTP protocol. If you prefer to use STDIO, it must be installed locally on your machine.

You'll find setup instructions for both protocols in the sections below.

⚠️ Security Note

Important: When running your own HTTP server instance (not using the remote https://mcp.laratranslate.com/v1), all connected clients share the same Lara API credentials configured via LARA_ACCESS_KEY_ID and LARA_ACCESS_KEY_SECRET environment

---