# AI Services

<!-- WARNING: THIS FILE WAS AUTOGENERATED! DO NOT EDIT! -->

AI Services are a suite of intelligence APIs for common AI workloads. You can use these APIs to process media data and create customized solutions. AI Services are part of the [Zoom Build platform](/docs/build/).

This documentation covers API concepts, authentication, processing modes, and endpoint references for each AI service to help you integrate these capabilities into your applications.

To quickly get started, visit **[Zoom.ai](https://zoom.ai/)** to explore pricing plans, try interactive demos, and set up your account.

## Scribe API

[Scribe API](/docs/ai-services/scribe/) provides speech-to-text transcription for audio. It supports real-time (Live mode), fast synchronous (Fast mode), and large-scale batch (Batch mode) transcript processing.

### Key features

-   **Large-scale processing**

    Transcribes up to 10,000 audio files in a single job, or 1,000 files from a manifest

-   **Smart formatting**

    Automatically adds timestamps, punctuation, and speaker separation

-   **Common audio formats**

    Transcribes WAV, M4A, MP3, MP4, and WebM files

-   **Speech understanding**

    Supports multilingual recognition, language detection, custom vocabulary, Text Polishing, and labeling each transcript item with the participant or source that spoke it

## Summarizer API

[Summarizer API](/docs/ai-services/summarizer/) offers conversation and meeting summarization for transcripts from any system. It supports fast synchronous summarization and large-scale batch processing.

### Key features

-   **Conversation-focused summarization**

    Summarizes meeting, call, chat, and other dialogue-based transcript content

-   **Structured outputs**

    Returns recap, action item, summary, or full summary results for workflows and automation

-   **Flexible processing**

    Supports inline text requests and storage-based batch jobs for transcript files

## Speech Summarization API

[Speech Summarization API](/docs/ai-services/speech-summarization/) transcribes an uploaded audio or video recording and streams a Markdown summary in one request.

### Key features

-   **Recording summaries**

    Returns a completed transcript followed by summary updates over Server-Sent Events

-   **Summary templates**

    Selects a template automatically or uses a predefined or custom template

-   **Transcript citations**

    Includes optional references to the transcript evidence behind the summary

## Translator API

[Translator API](/docs/ai-services/translator/) delivers machine translation for text across a broad range of languages. It supports fast synchronous translation and large-scale batch processing.

### Key features

-   **Fast translation**

    Translates plain text in a single synchronous API call

-   **Batch processing**

    Processes multiple files asynchronously for large-scale translation workloads

-   **Broad language support**

    Translates directly between any two supported languages, including all 24 official EU languages

## Decisions API

[Decisions API](/docs/ai-services/decisions/) evaluates shared text or image input against typed questions and returns structured results in one synchronous response.

### Key features

-   **Typed decisions**

    Returns a probability for a condition, a selected choice, or a score against ordered levels

-   **Multiple questions**

    Evaluates several independent questions against the same input in one request

-   **Text and image input**

    Accepts text, images, or both as evidence

## Voice Agent API

[Voice Agent API](/docs/ai-services/voice-agent/) provides real-time spoken conversation with transcription, synthesized speech, automatic turn detection, and function calling.

### Key features

-   **Bidirectional voice interaction**

    Streams user audio and agent responses over one connection

-   **Natural turn taking**

    Detects when the user finishes speaking and lets the user interrupt the agent

-   **Model choice**

    Uses the built-in model or your own OpenAI or Anthropic key

-   **Function calling**

    Connects spoken interactions to your application's tools and data