> ## Documentation Index
> Fetch the complete documentation index at: https://docs.corti.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Speech to Text implementation guides

> Build a full dictation experience on Corti Speech to Text: the client-side patterns beyond the raw API

The Corti Speech to Text API turns audio into transcript and command messages. Turning those messages into a real dictation *experience* (a microphone your users control, text that lands in the right field with the right spacing, spoken commands that drive your UI) is client-side work the API doesn't do for you.

These guides walk through that client implementation, one capability at a time. Each is paired with a **runnable example** in [`corti-examples`](https://github.com/corticph/corti-examples) and a short **video walkthrough**, so you can bootstrap your build from the example code or use it as reference.

***

## Studio vs. a client implementation

Before writing code, it's worth understanding where the [**Studio playground**](https://console.corti.app/) in the Corti Console fits.

Studio lets you exercise the **raw API** directly: configure a session, stream audio, and inspect the exact `transcript`, `command`, and status messages the server returns, then copy the corresponding code samples. It's the fastest way to validate a configuration and understand the API surface.

What Studio *doesn't* do is the client-side integration around those messages. That gap is exactly what these guides fill:

|                | Studio playground                         | Dictation guides + examples                         |
| :------------- | :---------------------------------------- | :-------------------------------------------------- |
| **Goal**       | Explore and validate the raw API          | Build a production client experience                |
| **Audio**      | Streams from your mic to show output      | Microphone capture *you* own and control            |
| **Transcript** | Displays returned messages                | Inserted at the caret with correct spacing & casing |
| **Commands**   | Shows `command` events fired              | Turned into real editor and app actions             |
| **Config**     | Try `replacements`, `keyterms`, etc. live | Managed, persisted, and shipped with your app       |
| **Best for**   | API exploration & code samples            | Reference and bootstrap for your own build          |

<Tip>
  Use Studio to confirm *what* the API returns for a given configuration, then use these guides to build *how* your application reacts to it.
</Tip>

***

## How the guides work

Every guide follows the same shape:

<Steps titleSize="h3">
  <Step title="Read the guide">
    A focused walkthrough of one capability, with the key code inline. Snippets are written to be portable — swap in your own auth and editor.
  </Step>

  <Step title="Watch the walkthrough">
    A short video demoing the matching example applet, embedded at the top of each guide.
  </Step>

  <Step title="Grab the example">
    The complete, runnable applet in `corti-examples`. Bootstrap your own build from it, or read it as a reference implementation.
  </Step>
</Steps>

<Note>
  The examples build on the [Dictation Web Component](/sdk/dictation/overview) wherever possible — it handles microphone management, streaming, and the transcript UI so the guides can focus on your application logic. The [Microphone](/stt/guides/dictation-microphone) guide shows the lower-level raw-SDK path for teams that need to own audio capture themselves.
</Note>

<iframe className="w-full aspect-video rounded-xl" src="https://www.youtube.com/embed/d6WKOaS9_Ts?start=0&end=66" title="Speech to Text Guides — overview walkthrough" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowFullScreen />

***

## What's covered

<Card title="Building a Dictation App" icon="mic" horizontal>
  Capture and insertion, recognition tuning, command-and-control, and audio handling:

  [Microphone](/stt/guides/dictation-microphone) · [Text Insertion](/stt/guides/dictation-text-insertion) · [Replacements](/stt/guides/dictation-replacements) · [Keyterms](/stt/guides/dictation-keyterms) · [Commands](/stt/guides/dictation-commands) · [App Control](/stt/guides/dictation-app-control) · [Dictation Box](/stt/guides/dictation-dictation-box) · [Device Buttons](/stt/guides/dictation-device-buttons) · [Audio Archive](/stt/guides/dictation-audio-archive) · [File Transcription](/stt/guides/dictation-file-transcription)
</Card>

<Card title="Ambient Documentation" icon="users" horizontal>
  Handle multi-speaker `/streams` output correctly:

  [Diarized Transcripts](/stt/guides/diarized-transcripts)
</Card>

<Card title="Augment STT with Agents" icon="zap" horizontal>
  Run Corti Agentic Framework agents over dictated text and finalized transcripts:

  [On-Demand Agent](/stt/guides/agent-on-demand) · [Second-Pass Agent](/stt/guides/agent-second-pass) · [Wake-Command Agent](/stt/guides/agent-wake-command) · [Conversational Agent](/stt/guides/agent-conversational)
</Card>

***

## Where to start

New to Corti dictation? Start with the [Dictation Web Component](/sdk/dictation/overview) for the fastest path to a working microphone and transcript, then follow the guides in order. If you need to integrate an existing microphone or build a fully custom UI, begin with [Microphone](/stt/guides/dictation-microphone).

<Note>Please [contact us](mailto:help@corti.ai) for help or questions.</Note>