OpenClaw

OpenClaw Talk Mode & Canvas👨‍💻

OpenClaw provides two interaction modes that go beyond text chat: Talk Mode for voice-driven conversations and Canvas for rich visual output. Talk Mode lets you speak to your agents and hear spoken responses, powered by speech-to-text (STT) for transcription and text-to-speech (TTS) for responses. Optional wake word detection activates Talk Mode from a background listener, so you never leave your keyboard. Canvas is a macOS-native surface where agents render HTML, charts, code diffs, and web pages through the A2UI (Agent-to-UI) protocol. Where text chat limits agents to sequential, text-only responses, Canvas gives them a visual output layer for dashboards, data visualizations, and interactive content. Together, Talk Mode and Canvas turn OpenClaw into a multimodal assistant that you can talk to, see output from, and interact with visually -- all while keeping your hands free. This concept page covers the configuration of both modes, the A2UI protocol, platform support, and practical workflows that combine voice and visual interaction.

Key Takeaways

  • 1Talk Mode enables real-time voice conversations with agents. Speech-to-text (Whisper) transcribes your voice input, the agent processes the transcribed text, and text-to-speech synthesizes a spoken response. The full cycle -- listen, transcribe, process, synthesize, play -- runs in a single continuous loop for natural conversational flow. The agent receives a standard text message regardless of input method, so your agent code does not need voice-specific logic. This means any existing agent works with Talk Mode without modification.
  • 2Voice wake detection runs as a background listener. Configure a wake word (e.g., "hey claw") and Talk Mode activates automatically when the word is spoken. This is supported on macOS (native) and Linux (via PulseAudio/ALSA). Windows, iOS, and Android are not supported for wake detection.
  • 3Canvas is a macOS-native UI surface that agents use to present visual content. Agents send A2UI (Agent-to-UI) commands to render HTML, navigate web pages, display charts, and show code diffs. The Canvas opens as a system overlay window alongside the chat. It supports any content that can be expressed as HTML and JavaScript, from a simple table to an interactive D3 visualization. This makes it the primary interface for tasks where text output alone is insufficient -- data analysis, visual QA, dashboard monitoring, and code review.
  • 4The A2UI (Agent-to-UI) protocol is the interface between agents and the Canvas surface. Agents send JSON commands with an action field (present for rendering content, navigate for opening URLs, screenshot for capturing page images) and a content payload. Any valid HTML, CSS, and JavaScript can be rendered in the Canvas. The protocol is unidirectional -- agents push content to the Canvas, and the Canvas renders it. User interactions within the Canvas (clicks, form submissions) are not sent back to the agent. For bidirectional interaction, use the chat or voice channel alongside Canvas.
  • 5Voice notes from messaging platforms (WhatsApp, Telegram) are automatically transcribed by the Gateway before reaching the agent. The agent receives plain text regardless of whether the user typed or sent a voice message -- transcription is transparent. This means agents that already work with text chat automatically support voice notes from mobile users without any configuration change. The Gateway handles the STT step in the message pipeline, before routing the normalized text to the target agent.
  • 6Talk Mode and Canvas work together for a fully hands-free workflow. Use Talk Mode for the conversation and Canvas for visual output -- verbally ask for a chart and see it rendered in the Canvas. A developer debugging a production issue can say "show me the error rate dashboard" and the agent navigates to the monitoring page on Canvas, takes a screenshot for analysis, and responds verbally with a summary. This combination is particularly effective for data analysis, deployment monitoring, and code review workflows where both voice and visual context improve the interaction.

Master openclaw talk mode & canvas

Take the Talk Mode & Voice course with hands-on lessons and challenges.

Examples

Configuring Talk Mode with wake word detection

json

Add this to your agent configuration in openclaw.json. The tts block configures text-to-speech (OpenAI TTS with the "nova" voice). The stt block configures speech-to-text (OpenAI Whisper for transcription). The wakeWord field sets the phrase that activates Talk Mode from the background listener. You can swap providers -- any OpenAI-compatible TTS/STT API works.

Starting Talk Mode and voice wake detection

bash

The talk command enters Talk Mode immediately -- the agent listens for voice, transcribes, processes, and responds with synthesized speech. The voicewake command manages the background listener that activates Talk Mode on the configured wake word. On macOS, a voice overlay appears with waveform visualization while Talk Mode is active.

Rendering visual content on Canvas with A2UI

json

This A2UI command renders an HTML report on the Canvas surface. The action field specifies the operation (present for rendering content). The content.type is html, and body contains the raw HTML to render. Agents generate these commands automatically when visual output is appropriate -- you do not need to construct them manually.

Navigating a web page on Canvas with screenshot

json

The navigate action opens a URL in the Canvas window. When screenshot is true, the agent receives a screenshot of the rendered page for visual analysis. This enables agents to browse the web, review pull requests, inspect dashboards, and report back on what they see -- all within the Canvas surface.

Voice-driven Canvas workflow (hands-free)

bash

This shows Talk Mode and Canvas working together. The voice wake listener activates Talk Mode when you speak the wake word. The agent receives your transcribed request, renders the dashboard on Canvas, and responds with synthesized speech. You see the visual output on the Canvas and hear the spoken summary -- entirely hands-free.

Common Mistakes

Mistake:

Using a short or common wake word like "hey" or "go", causing frequent false activations

Fix:

Choose a distinctive multi-word phrase with two or three syllables that does not commonly appear in conversation. Test it in your actual working environment with background noise to verify reliability.

Mistake:

Expecting Canvas to work on Linux or Windows -- Canvas is macOS-only

Fix:

The Canvas surface requires macOS. On Linux and Windows, use the TUI (terminal interface) or WebChat for agent interaction. Voice wake detection works on macOS and Linux, but the Canvas visual surface is exclusive to macOS.

Mistake:

Using a large or slow STT model for real-time voice interaction, causing noticeable transcription delay

Fix:

Use OpenAI's whisper-1 or equivalent fast models for real-time interaction. Larger models may be more accurate but introduce latency that disrupts conversational flow. Prioritize speed for Talk Mode; use higher-accuracy models only for batch transcription of recorded audio.

Mistake:

Rendering heavy JavaScript applications or full web apps in Canvas, consuming excessive memory

Fix:

Keep Canvas content focused and lightweight. Render targeted HTML -- a chart, a table, a code diff -- rather than embedding entire web applications. The Canvas is a presentation surface, not a browser replacement.

Mistake:

Writing agent code that handles voice input differently from text input

Fix:

Voice input is transcribed to text before it reaches the agent. The agent receives a standard text message regardless of whether the user typed or spoke. Do not add voice-specific code paths in your agent logic -- the Gateway handles transcription transparently.

Best Practices

  • Start with Talk Mode before adding Canvas. Get voice interaction working first -- configure STT/TTS, test the wake word, verify transcription accuracy -- then layer Canvas on top for visual output.
  • Use distinctive, multi-syllable wake words and test them in your actual environment. Background noise, microphone quality, and room acoustics all affect detection accuracy.
  • Combine Talk Mode and Canvas for hands-free visual workflows. Verbally request a chart, dashboard, or code review and see the result rendered on Canvas while hearing a spoken summary.
  • Keep Canvas content lightweight and focused. Render a single chart, table, or code block rather than a full application. The Canvas is strongest as a targeted visual display, not a general-purpose browser.
  • Use the A2UI screenshot capability for visual analysis tasks. Have the agent navigate to a URL, take a screenshot, and analyze the page content -- this is particularly effective for reviewing pull requests, monitoring dashboards, and checking deployment status.
  • Test voice note transcription from messaging platforms. Send a WhatsApp or Telegram voice note to your agent and verify the transcription quality and response accuracy before relying on it for production workflows.

Summary

OpenClaw Talk Mode and Canvas extend agent interaction beyond text into voice and visual modalities. Talk Mode uses STT (Whisper) and TTS (OpenAI) for spoken conversations, with wake word detection on macOS and Linux for hands-free activation. Canvas is a macOS-native surface where agents render HTML, charts, code diffs, and web pages through the A2UI protocol. Voice notes from WhatsApp and Telegram are transcribed transparently by the Gateway before reaching the agent. The two modes combine into a hands-free visual workflow: speak a request via Talk Mode, see the result rendered on Canvas, and hear a spoken summary. Platform support differs -- Talk Mode works on macOS and Linux, while Canvas is macOS-only. On non-macOS platforms, the TUI and WebChat provide alternative interaction surfaces.

Practice OpenClaw with hands-on challenges

Learn openclaw talk mode & canvas hands-on in your IDE

Interactive lessons and challenges on Stanza, practice in VS Code, Cursor, or the web.

Related Concepts

Related Cheatsheets

Master OpenClaw with Stanza

Interactive lessons and challenges, right in your code editor.

Check the free courses. No credit card.