# How to build with Gemini 3.5 Transcribe

## Executive summary

Google DeepMind launched Gemini 3.5 Transcribe, an LLM-based transcription model available via both the Interactions API and the Live API. This model significantly enhances accuracy by correctly transcribing complex data types—such as email addresses, phone numbers, and mixed units of measurement—and maintaining high performance across over 85 supported languages, even when language codes are set to English.

## Key takeaways

- LLM-Based Transcription Model: The model's LLM foundation allows it to handle complex data structures and context better than traditional transcription models. For example, it can correctly identify and edit email addresses even if spoken phonetically (e.g., 'tosten at google.com').
- Handling Complex Data Types: Gemini 3.5 Transcribe accurately recognizes specific formats, including US phone numbers and international variations (e.g., Singapore's 8-digit format). It can also correctly interpret units of measure (e.g., meters vs. centimeters).
- Multilingual and Customization Support: The model supports over 85 languages, automatically recognizing spoken language even if language hints are set to English. Accuracy can be further improved by providing custom vocabulary (e.g., names of people in a meeting) or setting specific language codes.

## Technical details

- Model Availability: Gemini 3.5 Transcribe is available for both unary transcription via the Interactions API and real-time streaming via the Live API.
- API Integration: The model's deep understanding of language and reasoning capabilities make it suitable for new applications requiring highly accurate, real-time transcription.

## Practical implications

- Build engineers can integrate Gemini 3.5 Transcribe into applications requiring high-fidelity speech-to-text conversion for specialized data (e.g., medical dictation, financial call centers).
- The dual API availability (Interactions/Live) allows developers to build both batch processing and real-time streaming features.
- Custom vocabulary integration enables domain-specific improvements in transcription accuracy.

## Topics

Speech Recognition, Large Language Models (LLMs), API Integration, Natural Language Processing (NLP), Google Developers, Interactions API, Live API

Source: https://www.youtube.com/watch?v=PWPsZssgfDo