Hooks

React hooks from the Pipecat React SDK for client access, media device management, and bot event handling, like usePipecatClient.

The Pipecat React SDK provides hooks for accessing client functionality, managing media devices, and handling events.

​

usePipecatClient

Provides access to the PipecatClient instance originally passed to PipecatClientProvider. Returns PipecatClient | undefined, which is undefined outside a provider.

import { usePipecatClient } from "@pipecat-ai/client-react";

function StartButton() {
  const pcClient = usePipecatClient();

  const start = async () => {
    await pcClient?.startBotAndConnect({
      endpoint: "/api/start",
      requestData: {
        // Any custom data your /start endpoint requires
      },
    });
  };

  return <button onClick={start}>Start</button>;
}

​

useRTVIClientEvent

Allows subscribing to RTVI client events. It is advised to wrap handlers with useCallback.

import { useCallback } from "react";
import { RTVIEvent, TransportState } from "@pipecat-ai/client-js";
import { useRTVIClientEvent } from "@pipecat-ai/client-react";

function EventListener() {
  useRTVIClientEvent(
    RTVIEvent.TransportStateChanged,
    useCallback((transportState: TransportState) => {
      console.log("Transport state changed to", transportState);
    }, [])
  );
}

Arguments

event · RTVIEvent · required

handler · function · required

​

usePipecatClientMediaDevices

Manage and list available media devices.

import { usePipecatClientMediaDevices } from "@pipecat-ai/client-react";

function DeviceSelector() {
  const {
    availableCams,
    availableMics,
    availableSpeakers,
    selectedCam,
    selectedMic,
    selectedSpeaker,
    updateCam,
    updateMic,
    updateSpeaker,
  } = usePipecatClientMediaDevices();

  return (
    <>
      <select
        name="cam"
        onChange={(ev) => updateCam(ev.target.value)}
        value={selectedCam?.deviceId}
      >
        {availableCams.map((cam) => (
          <option key={cam.deviceId} value={cam.deviceId}>
            {cam.label}
          </option>
        ))}
      </select>
      <select
        name="mic"
        onChange={(ev) => updateMic(ev.target.value)}
        value={selectedMic?.deviceId}
      >
        {availableMics.map((mic) => (
          <option key={mic.deviceId} value={mic.deviceId}>
            {mic.label}
          </option>
        ))}
      </select>
    </>
  );
}

​

useMediaState

Subscribes to the per-device mic and cam state, re-rendering when either transitions. Use it to drive permission UI.

import { useMediaState } from "@pipecat-ai/client-react";

function MicStatus() {
  const { mic } = useMediaState();

  if (mic.state === "error") {
    return <p>Microphone unavailable: {mic.reason}</p>;
  }
  return <p>Microphone {mic.state}</p>;
}

Returns a MediaState snapshot. Safe to call from any component inside PipecatClientProvider: the subscription is mounted for you.

​

useMediaSupport

Subscribes to which kinds of media can flow in the session, re-rendering when that changes. Use it to hide controls for media the transport or bot doesn’t support.

import { useMediaSupport } from "@pipecat-ai/client-react";

function CamControl() {
  const { cam } = useMediaSupport();

  if (cam === false) return null;
  return <CamToggle />;
}

Returns a MediaSupport snapshot. Safe to call from any component inside PipecatClientProvider: the subscription is mounted for you.

​

useBotCapabilities

Returns the bot’s BotCapabilities from its bot-ready message, or undefined until the bot is ready, after disconnecting, and for bots that don’t send capabilities.

import { useBotCapabilities } from "@pipecat-ai/client-react";

function MetricsPanel() {
  const capabilities = useBotCapabilities();

  if (capabilities?.metrics === false) return null;
  return <Metrics />;
}

Safe to call from any component inside PipecatClientProvider: the subscription is mounted for you.

​

usePipecatClientMediaTrack

Access audio and video tracks.

import { usePipecatClientMediaTrack } from "@pipecat-ai/client-react";

function MyTracks() {
  const localAudioTrack = usePipecatClientMediaTrack("audio", "local");
  const botAudioTrack = usePipecatClientMediaTrack("audio", "bot");
}

Arguments

trackType · 'audio' | 'video' | 'screenAudio' | 'screenVideo' · required

participantType · 'bot' | 'local' · required

Returns the MediaStreamTrack, or null until one has started.

​

usePipecatClientAudioElement

Returns a RefObject<HTMLAudioElement> to attach to your own <audio> element, which then plays the bot’s audio track. It is what PipecatClientAudio renders: it retries playback on the user’s first click, tap, or keypress when the browser blocks autoplay, and follows speaker changes with setSinkId.

import { usePipecatClientAudioElement } from "@pipecat-ai/client-react";

function BotAudio() {
  const audioRef = usePipecatClientAudioElement();
  return <audio ref={audioRef} autoPlay />;
}

​

usePipecatClientScreenShareControl

Controls the local participant’s screen share state.

import { usePipecatClientScreenShareControl } from "@pipecat-ai/client-react";

function ScreenShareToggle() {
  const { enableScreenShare, isScreenShareEnabled } =
    usePipecatClientScreenShareControl();

  return (
    <button onClick={() => enableScreenShare(!isScreenShareEnabled)}>
      {isScreenShareEnabled ? "Stop Sharing" : "Share Screen"}
    </button>
  );
}

Starting a share opens the browser’s own picker, so call enableScreenShare(true) from a user gesture. Render the result with PipecatClientVideo using trackType="screenVideo", or reach the raw track through tracks().

​

usePipecatClientTransportState

Returns the current transport state.

import { usePipecatClientTransportState } from "@pipecat-ai/client-react";

function ConnectionStatus() {
  const transportState = usePipecatClientTransportState();
}

​

usePipecatClientCamControl

Controls the local participant’s camera state.

import { usePipecatClientCamControl } from "@pipecat-ai/client-react";
function CamToggle() {
  const { enableCam, isCamEnabled } = usePipecatClientCamControl();

  return (
    <button onClick={() => enableCam(!isCamEnabled)}>
      {isCamEnabled ? "Disable Camera" : "Enable Camera"}
    </button>
  );
}

​

usePipecatClientMicControl

Controls the local participant’s microphone state.

import { usePipecatClientMicControl } from "@pipecat-ai/client-react";
function MicToggle() {
  const { enableMic, isMicEnabled } = usePipecatClientMicControl();

  return (
    <button onClick={() => enableMic(!isMicEnabled)}>
      {isMicEnabled ? "Disable Microphone" : "Enable Microphone"}
    </button>
  );
}

​

useDTMF

Sends DTMF keypresses to the bot: the digits a caller would press on a phone keypad.

import { useDTMF } from "@pipecat-ai/client-react";

function Keypad() {
  const { sendTone } = useDTMF();

  return (
    <div>
      {["1", "2", "3", "4", "5", "6", "7", "8", "9", "*", "0", "#"].map(
        (key) => (
          <button key={key} onClick={() => sendTone(key)}>
            {key}
          </button>
        ),
      )}
    </div>
  );
}

sendTone takes a single key or a sequence ("123#"), and wraps sendDTMF(), so the same validation and protocol requirements apply. It is a no-op until a client is available from PipecatClientProvider.

​

usePipecatConversation

The primary hook for accessing the conversation message stream. Returns the current list of messages (ordered for display) and a function to inject messages programmatically.

Each assistant message’s text parts are split into spoken and unspoken segments based on real-time speech progress, so you can style them differently (e.g. dim unspoken text).

import { usePipecatConversation } from "@pipecat-ai/client-react";
import type {
  BotOutputText,
  ConversationMessage,
} from "@pipecat-ai/client-react";

function Messages() {
  const { messages } = usePipecatConversation({
    onMessageCreated(message: ConversationMessage) {
      console.log("New message:", message);
    },
    onMessageUpdated(message: ConversationMessage) {
      if (message.final) {
        console.log("Message finalized:", message);
      }
    },
  });

  return (
    <ul>
      {messages.map((msg, i) => (
        <li key={`${msg.createdAt}-${i}`}>
          <strong>{msg.role}:</strong>{" "}
          {msg.parts?.map((part, j) => {
            if (typeof part.text === "string") {
              return <span key={j}>{part.text}</span>;
            }
            const { spoken, unspoken } = part.text as BotOutputText;
            return (
              <span key={j}>
                <span>{spoken}</span>
                <span style={{ opacity: 0.5 }}>{unspoken}</span>
              </span>
            );
          })}
        </li>
      ))}
    </ul>
  );
}

Options

onMessageCreated · (message: ConversationMessage) => void

Called once when a new message first enters the conversation. The message may or may not be complete at this point; check message.final.

onMessageUpdated · (message: ConversationMessage) => void

Called whenever an existing message’s content changes (e.g. streaming text appended, function call status changed, message finalized). Check message.final to detect finalization.

outputTypeMetadata · Record<string, OutputTypeMetadata>

Metadata for output types to control rendering and speech progress behavior. Used to determine which output types should be excluded from position-based speech splitting.

aggregationMetadata · Record<string, OutputTypeMetadata> · deprecated

Deprecated. Use outputTypeMetadata instead. Where both have an entry for an output type, the outputTypeMetadata entry is used.

botOutputFilter · BotOutputFilter

Restrict which bot output reaches messages, as { spoken?: boolean; unspoken?: boolean }. Use it to render only what the bot says aloud, or only the text it produces without speaking.

includeBackchannel · boolean · default: "false"

Include backchannels: short acknowledgments said while the other side talks, such as the user’s “mhm” or the bot’s “Mm-hmm.” (RTVI 2.2.0+). Each is its own message with backchannel: true, placed by when it was said, and is also reported to onMessageCreated.

Returns

messages · ConversationMessage[]

The current list of conversation messages, ordered for display. Assistant messages have their text parts split into { spoken, unspoken } based on real-time speech progress.

injectMessage · (message: { role: 'user' | 'assistant' | 'system'; parts: ConversationMessagePart[] }) => void

Programmatically inject a message into the conversation (e.g. a system prompt or user-typed input).

sendText · (text: string, options?: SendTextOptions) => Promise<string>

Sends text to the bot and adds it to the conversation as a user message. The message’s status is sent, then received once the bot acknowledges it, or failed if it couldn’t be sent. Bots before RTVI 2.2.0, or with user input turned off, never acknowledge text, so their messages stay sent. Use it in place of injectMessage plus client.sendText(), not alongside them. Resolves to the ID of the send-text message, also the message’s msgId.

botOutputEvents · BotOutputEvent[]

The raw bot-output events behind messages, in arrival order. Useful for debugging or replaying how a response was assembled.

​

useConversationContext

Lower-level hook that provides direct access to the conversation context. Use this when you only need injectMessage without subscribing to the message stream, or to check whether the connected bot supports BotOutput events.

import { useConversationContext } from "@pipecat-ai/client-react";

function TextInput() {
  const { injectMessage, botOutputSupported } = useConversationContext();

  const send = (text: string) => {
    injectMessage({
      role: "user",
      parts: [{ text, final: true, createdAt: new Date().toISOString() }],
    });
  };

  return (
    <input
      onKeyDown={(e) => e.key === "Enter" && send(e.currentTarget.value)}
    />
  );
}

Returns

injectMessage · (message: { role: 'user' | 'assistant' | 'system'; parts: ConversationMessagePart[] }) => void

Programmatically inject a message into the conversation.

sendText · (text: string, options?: SendTextOptions) => Promise<string>

Sends text to the bot and adds it to the conversation as a user message with a status of sent, received or failed.

botOutputSupported · boolean | null

Whether the connected bot supports BotOutput events (RTVI 1.1.0+). null means detection hasn’t completed yet.

​

Conversation Utilities

Helpers the conversation store uses, exported for building your own message lists from ConversationMessage[]:

Function	Signature	Description
sortByCreatedAt	(a: ConversationMessage, b: ConversationMessage) => number	Comparator that orders messages by createdAt.
isMessageEmpty	(message: ConversationMessage) => boolean	Whether a message has no text. Function call messages are never empty.
filterEmptyMessages	(messages: ConversationMessage[]) => ConversationMessage[]	Drops empty messages, keeping one only if no later non-empty message has the same role.
mergeMessages	(messages: ConversationMessage[]) => ConversationMessage[]	Merges a user or assistant message into the previous one of the same role while that one isn’t final and is less than 30 seconds older. Text sent with sendText() keeps its own message.
deduplicateFunctionCalls	(messages: ConversationMessage[]) => ConversationMessage[]	Keeps one function call message per tool_call_id, the one with the most advanced status.
filterBotOutputText	(text: BotOutputText, filter?: BotOutputFilter) => BotOutputText	Applies a BotOutputFilter, emptying the spoken or unspoken portion whose flag is false.

​

User Interface

These hooks voice-enable a UI: stream the screen to a server-side UIWorker, send UI events, handle the bot’s UI commands, and track background work. They must be used within a PipecatClientProvider. See the RTVI standard for wire payloads and Controlling the UI for the patterns.

​

useUISnapshot

Streams accessibility snapshots of the page to the bot. Call once near the root of your app.

useUISnapshot({ enabled: true, debounceMs: 300, trackViewport: true });

Arguments: options?: UseUISnapshotOptions with enabled (default true), debounceMs (default 300), trackViewport (default true), and logSnapshots (default false).

​

useUIEventSender

Returns (event: string, payload?) => void to send app-defined UI events to the bot (routed to the worker’s @ui_event handlers).

const sendUIEvent = useUIEventSender();
sendUIEvent("note_click", { ref: "e8" });

​

useUICommandHandler

Registers a handler for a named UI command, active while the component is mounted.

useUICommandHandler("add_note", (payload) => addNoteToPage(payload));

Arguments

command · string · required

handler · (payload: T) => void | Promise<void> · required

​

Default Command Handlers

useDefaultUICommandHandlers(options?) installs DOM handlers for all the standard commands at once (scroll_to, focus, highlight, select_text, set_input_value, click). Each resolves the target by snapshot ref, then target_id, and refuses unsafe targets. Install them individually with useDefaultScrollToHandler, useDefaultFocusHandler, useDefaultHighlightHandler, useDefaultSelectTextHandler, useDefaultSetInputValueHandler, and useDefaultClickHandler. For the non-DOM commands, useToastHandler(handler) and useNavigateHandler(handler) are typed convenience wrappers.

useUISnapshot();
useDefaultUICommandHandlers();

Options: useDefaultUICommandHandlers takes one options object per handler, and each individual hook takes the matching object.

scrollTo · DefaultScrollToOptions

Show DefaultScrollToOptions

block · ScrollLogicalPosition · default: "'start'"

scrollIntoView block position.

inline · ScrollLogicalPosition · default: "'nearest'"

scrollIntoView inline position.

defaultBehavior · ScrollBehavior · default: "'smooth'"

Scroll behavior when the command’s behavior is unset.

container · Element | null | (() => Element | null | undefined)

Scroll inside this element instead of the nearest scrollable ancestor. The function form is evaluated on each scroll.

offset · { top?: number; left?: number }

Pixel offsets applied after scrolling, such as to clear a sticky header. Applied only when container is set.

focus · DefaultFocusOptions

Show DefaultFocusOptions

preventScroll · boolean · default: "false"

Pass { preventScroll: true } to element.focus() so focusing doesn’t scroll the page.

highlight · DefaultHighlightOptions

Show DefaultHighlightOptions

className · string · default: "'ui-highlight'"

CSS class toggled on the target for the highlight’s duration. Your app styles it.

defaultDurationMs · number · default: "1500"

Duration when the command’s duration_ms is missing.

scrollIntoViewFirst · boolean · default: "false"

Scroll the target into view before highlighting it.

selectText · DefaultSelectTextOptions

Show DefaultSelectTextOptions

scrollIntoViewFirst · boolean · default: "true"

Scroll the target into view before selecting.

block · ScrollLogicalPosition · default: "'center'"

scrollIntoView block position when scrolling first.

setInputValue · DefaultSetInputValueOptions

Show DefaultSetInputValueOptions

focusFirst · boolean · default: "false"

Focus the field before writing, then blur it after the change events fire.

​

useUIJobGroups

Reads accumulated job groups (oldest first) and controls from the nearest UIJobGroupsProvider. Without one, it returns no groups and its controls do nothing. Wrap your app in the provider to collect them:

<UIJobGroupsProvider maxGroups={20}>
  <App />
</UIJobGroupsProvider>

const { groups, cancelJobGroup, dismissJobGroup, clearCompleted } =
  useUIJobGroups();

Returns: UIJobGroupsAPI: groups: JobGroup[], cancelJobGroup(jobId, reason?), dismissJobGroup(jobId), and clearCompleted(). Each JobGroup has jobId, label, cancellable, status, startedAt, completedAt, and jobs (per-worker { workerName, status, startedAt, completedAt, updates, response }).