> For the complete documentation index, see [llms.txt](https://blockgram.gitbook.io/blockgram-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://blockgram.gitbook.io/blockgram-docs/about-blockgram/core-features/real-time-translation-engine.md).

# Real-Time Translation Engine

BlockGram breaks language barriers instantly, right inside your Telegram chats. Users can communicate seamlessly across languages, whether coordinating globally or managing international communities, without leaving the chat interface.

#### Real-Time Language Detection and Translation

Every message is automatically scanned using AI-powered language detection. The system evaluates character usage, syntax, message history, and Telegram locale settings to determine the language.

**Translation Triggers**

* **Manual:** Tap a message or reply with `/translate`
* **Automatic:** Messages in other languages are translated based on user preferences
* **Conditional:** Group-level rules can auto-translate specific messages, such as admin posts

Supported content includes text messages, voice messages (converted to text), message replies, forwards, inline queries, and bot messages.

#### AI Translation Engine

BlockGram uses transformer-based multilingual models fine-tuned for chat environments, preserving tone, informal expressions, and idioms.

| Feature          | Details                                                                                             |
| ---------------- | --------------------------------------------------------------------------------------------------- |
| Language Support | 100+ languages, including English, Spanish, Arabic, Hindi, Russian, Mandarin, and regional dialects |
| Model Features   | Context-aware, maintains idioms, gender, and tone                                                   |
| Accuracy         | Near-human, 90–95% fidelity for casual conversation                                                 |

**Multi-Turn Awareness**\
The system tracks conversation context across multiple messages to maintain consistent translations, handling pronouns, references, and ongoing questions.

#### User Interface and Experience

**Inline Previews**

* Tap a message to see “Translate to \[your language]”
* Toggle between original and translated text
* Translations appear directly under the original, not as separate messages

**Voice Messages**

* Real-time transcription and translation using speech-to-text models
* Playback or display of translated audio/text

**Smart Suggestions**

* The AI detects foreign-language messages and suggests translation with one tap

#### Group and Community Tools

**Global Community Management**

* Admins can set default translation languages per user or group
* Auto-translate rules ensure consistent communication
* Multi-language broadcasts allow one message to reach all users in their preferred language

**Multilingual Content**

* Polls, bot commands, and links are automatically localized

#### Performance and Architecture

* Frequently repeated phrases are cached using semantic embeddings to reduce latency
* Lightweight translation models run on-device, while complex translations are handled via encrypted cloud servers
* Edge and cloud hybrid design ensures speed, privacy, and accuracy

#### Privacy and Control

* Translations are visible only to the user by default
* No data is stored after translation unless explicitly saved
* Translation can be disabled per chat or per user

#### Example Use Cases

* **Global DAO Chat:** A Mandarin message is instantly translated to English while keeping the original for reference
* **Voice Updates:** A Spanish voice note is transcribed and translated in real time
* **Community Announcements:** One admin message is automatically delivered in multiple languages to all members

#### Why It Matters

BlockGram doesn’t just translate words. It preserves context, intent, and connection. Real-time multilingual support ensures every participant, no matter the language, can fully engage in Web3 conversations.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://blockgram.gitbook.io/blockgram-docs/about-blockgram/core-features/real-time-translation-engine.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.