Claude Gets a Voice Upgrade: Pick Your Model, Connect Your Apps
Anthropic has updated Claude's voice mode so it can use smarter AI models and reach into Gmail, Slack and Notion, going further than OpenAI's voice feature currently does.
Key points
- Anthropic updated Claude's voice mode on Thursday to support its Opus, Sonnet and Haiku AI models, not just the basic Haiku.
- The new voice mode can connect to external apps including Gmail, Google Calendar, Slack, Canva and Notion.
- Free users get voice mode in beta but are limited to Haiku with only one connected app.
- Claude supports ten languages in voice mode, though users must set the language manually.
- Anthropic didn't upgrade the underlying voice technology, so conversational feel and interruption handling stay the same.
Talk to your phone, draft an email, shift a meeting, start a Notion document, all without touching a screen. That's the pitch behind Claude's refreshed voice mode, which Anthropic rolled out Thursday.
What actually changed?
The biggest upgrade is model choice. Until now, Claude's voice mode only ran on Haiku, the fastest but least powerful of Anthropic's three AI model tiers. Think of the tiers like car engines: Haiku is a nippy city engine, Sonnet is a mid-range tourer, and Opus is the full-power option. Quick answers were fine on Haiku, but anything complex stretched it thin.
Now voice mode picks whichever model you last used in regular text chat, then runs its fastest version of that model. More brainpower, no settings fiddling.
The second change is bigger in practical terms. Claude's voice can now reach into outside apps: Gmail, Google Calendar, Slack, Canva (the design tool) and Notion (the note-taking and project app). Reschedule Thursday's meeting, fire off a draft message, start a document. As first reported by TechCrunch AI, this puts Claude ahead of OpenAI's voice mode, which updated its conversational style but still can't touch other tools to get work done. We noted a similar pattern when Google's AI Mode gained app connections on 16 July, linking Canva and others directly to its conversational search tool.
Who gets what?
| User tier | Models available | Connected apps |
|---|---|---|
| Free | Haiku only | 1 app |
| Paid (all tiers) | Haiku, Sonnet, Opus | Multiple apps |
All users get the update in beta across platforms. Free users are capped at Haiku and one connected app, so the full benefit lands mainly with paying subscribers.
Anthropics multilingual support, added earlier this year, now covers ten languages across Europe and Asia, from French and German to Hindi, Japanese and Korean, plus Brazilian Portuguese and two Spanish variants. One catch: you have to tell Claude which language to use. It won't detect it automatically.
Should you expect it to sound better?
Not yet. Anthropic didn't change the underlying voice technology, the software layer that controls tone and pacing, including how well it handles you jumping in mid-sentence. OpenAI recently overhauled exactly those qualities in ChatGPT's voice mode. Claude may sound identical to before, even with the smarter model underneath.
If you want a voice assistant that takes action across your work apps, the new Claude is worth trying. For natural back-and-forth conversation that feels closer to talking to a person, OpenAI still has an edge there for now. The real test will be whether Anthropic tightens the voice layer before competitors add tool access of their own.
Common questions
Do I need to pay to use Claude's voice mode?
No. Voice mode is free in beta, but free accounts are capped at the basic Haiku model and can only connect one outside app, such as Gmail or Slack.
Will Claude automatically detect what language I am speaking?
Not at the moment. You need to tell Claude which of its ten supported languages to use before you start talking.
