# Speech-to-Text overview (/en/realtime-media/speech-to-text)

> For AI agents: see the complete documentation index at [llms.txt](/llms.txt).

Agora's Real-Time Speech to Text transcribes live audio streams into text, enabling closed captions, live transcription, and AI-powered workflows. Translate transcribed text into multiple languages in real-time, or pass it directly to large language models to bridge real-time communication with intelligent applications.

## Quick links [#quick-links]

<Cards>
  <Card title="REST quickstart" href="./get-started/quickstart" description="Start, manage, query, and stop a real-time speech-to-text agent." />

  <Card title="REST API reference" href="/en/api-reference/api-ref/speech-to-text" description="Explore Real-Time Speech to Text API operations and parameters." />

  <Card title="Sample server" href="https://github.com/AgoraIO/agora-rtt-server" description="Browse the open-source Real-Time Speech to Text sample server." />
</Cards>

## Key features [#key-features]

<CardGrid>
  <FeatureCard title="Live transcription for RTC">
    Integrated with Agora’s voice and video service, live transcription and captions improve accessibility for your audience. Perfect for meetings, live streaming, lectures, interviews, live shopping, and more.
  </FeatureCard>

  <FeatureCard title="Real-time translation (Beta)">
    Break down language barriers with live speech-to-text translation to multiple languages during real-time communication or live streaming. The high accuracy translation text, delivered with ultra low latency, can be integrated with LLMs for enhanced capabilities.
  </FeatureCard>

  <FeatureCard title="Cloud-based STT">
    Cloud-based service converts voice to text for active or specific hosts and then distributes the text to all participants in the channel for further processing. The service does not depend on the client's device performance and network conditions.
  </FeatureCard>

  <FeatureCard title="Speaker labeling">
    Label each transcribed text with the speaker's UID. Separate transcription of each host ensures accuracy even when multiple hosts are talking simultaneously.
  </FeatureCard>

  <FeatureCard title="Caption recording">
    Upload the transcriptions as .vtt files to cloud storage, then play back audio or video recordings with closed captions (CC). The timestamps in the .vtt file ensure that the text is perfectly synchronized with the audio or video, so that it appears exactly where it was generated.
  </FeatureCard>

  <FeatureCard title="Multi-language support">
    Real-time transcription supports all major languages and dialects. Real-time translation supports translation into 40+ target languages.
  </FeatureCard>
</CardGrid>