Core Blocks
◆Block
Translate Block
Translate text between languages using AI
Translate Block
The Translate block converts text from one language to another using an LLM. It supports any language pair and preserves meaning, tone, and context better than rule-based translation engines.
Overview
| Property | Value |
|---|---|
| Type | translate |
| Category | Core Block |
| Color | #06B6D4 (Cyan) |
When to Use
- Translate user input or AI output to another language
- Build multilingual chatbots
- Localize content for different markets
- Translate support tickets before processing
Configuration
| Setting | Type | Description |
|---|---|---|
| Content | Long text | Text to translate: {{agent.content}} |
| Source Language | Dropdown | Input language (or Auto-Detect) |
| Target Language | Dropdown | Desired output language |
| Model | Dropdown | LLM for translation |
| API Key | Password | Provider API key |
| Preserve Formatting | Toggle | Keep markdown, HTML, or code formatting |
Outputs
| Field | Type | Description |
|---|---|---|
content | string | Translated text |
sourceLanguage | string | Detected source language |
targetLanguage | string | Target language used |
Example: Multilingual Support Bot
Goal: Accept support tickets in any language and respond in the user's language.
Workflow:
[Starter] → [Translate: to English] → [Agent: Support] → [Translate: to Source] → [Response]How it works:
- Customer writes in Spanish: "¿Cómo cambio mi contraseña?"
- First Translate block converts to English: "How do I change my password?"
- Agent processes the English question with its knowledge base
- Second Translate block converts the response back to Spanish
- Customer receives the answer in their own language
Tips
- Auto-Detect source language when you don't know what language users will use
- LLM translation handles idioms and context better than Google Translate for complex text
- Use GPT-4o or Claude for high-quality translations; mini models work for simple text