Why dictation isn't working in your browser

This is part of the answers hub. If the dictate button does nothing, the cause is almost always which browser you're in. Voice recognition support varies a lot across browsers.

The support gap

Dictation uses the Web Speech API's SpeechRecognition interface. Support is uneven:

• Chrome and Edge: supported and reliable (under the webkitSpeechRecognition prefix).
• Safari: partial support on recent versions, also prefixed; behavior is less consistent.
• Firefox — off by default. Recognition sits behind the dom.webspeech.recognition.enable flag in about:config, so for end users the button has nothing to call out of the box.

The full matrix is on the speech-recognition browser support reference.

Fix: if you're in Firefox, switch to Chrome or Edge for voice, or just type: typed notes produce the same async feedback document.

If you're in a supported browser

When recognition is available but still silent, check, in order:

1. Microphone permission. Click the mic icon by the address bar and allow the site. A blocked mic gives you a dead button.
2. The right input device. In your OS sound settings, confirm the active microphone is the one you're speaking into.
3. Exclusive access. Conferencing apps sometimes hold the mic. Close them and retry.
4. Your connection. Most recognition streams audio to a server, so a dropped connection stops results mid-sentence. Toggle the mic to restart the session.

When voice isn't the bottleneck

Dictation is a convenience. The artifact your agent reads is the same whether you spoke or typed it. If the real problem is the agent not using your input, that's a handoff-format issue, not a dictation one.

Frequently asked questions

Why is the dictate button missing or disabled in Firefox?

Firefox keeps the SpeechRecognition half of the Web Speech API disabled by default. It sits behind the dom.webspeech.recognition.enable flag in about:config, so for end users it isn't available out of the box. Type your notes instead, or switch to Chrome or Edge for voice.

Dictation worked, then stopped mid-session. Why?

Recognition sessions time out and the API can drop on network interruptions, since most implementations send audio to a cloud service. Toggle the mic off and on to start a fresh session. A flaky connection is the most common cause of a mid-session stop.

The browser isn't picking up my voice at all.

Check microphone permission for the site (the mic icon by the address bar), confirm the right input device is selected in your OS sound settings, and make sure no other app has exclusive hold of the mic.

Does dictation need an internet connection?

Usually yes. Most browser speech recognition streams audio to a server for transcription, so offline or on a dropped connection it won't return results.