Claude Apps Switch to Auto Mode by Default

Anthropic will make auto mode the default for sessions in its Claude web, desktop and mobile apps starting next week, the company said Thursday, Oct. 22. The change covers Pro, Max and Team subscribers and replaces step-by-step approval prompts with an automated check that blocks risky actions.

Key takeaways

• Sessions in Claude's web, desktop and mobile apps on Pro, Max and Team plans will run in auto mode by default starting next week; Anthropic has not named an exact date.
• Auto mode routes actions through a classifier instead of asking the user to approve each step. Users can switch back to manual approval at any time.
• The move extends an Aug. 14 change that made auto mode the default for new Claude Code sessions on the same plans.

What changes with Claude auto mode in the apps

In a post on its official X account, Anthropic said: "Starting next week, sessions across Claude web, desktop, and mobile apps will run in auto mode by default on Pro, Max, and Team plans. Rather than asking you to approve each step, an automated check blocks risky actions. Switch back to manual anytime."

The post did not give a specific start date or list which in-app features count as sessions. It named only the Pro, Max and Team plans.

How the auto mode classifier works

Auto mode first appeared in Claude Code in March and became generally available to all Claude Code users on July 10, according to Anthropic's auto mode explainer. Before each tool call runs, a classifier screens for actions such as "mass deleting files," "sensitive data exfiltration" and "malicious code execution."

When the classifier blocks an action, Claude usually looks for a safer route or asks the user. In Claude Code, the session falls back to manual approvals after three blocks in a row or 20 across a session, Anthropic said in August.

Why Anthropic says approval prompts fall short

Anthropic's Aug. 7 blog post laid out the case. Users approve 97% of permission prompts, the company said. In a controlled study with 1,053 testers, "human review caught just 13.6% of dangerous commands, while auto mode caught 89%."

In that study, auto mode blocked 800 commands that humans approved, while humans blocked six that auto mode allowed. In real sessions, 6.3% of manually approved sessions contained an unrequested harmful action rated severity 7 or higher, versus 2.4% of auto mode sessions.

Anthropic also cited red-teaming by Apollo Research that cut the classifier's miss rate from 12% to 7%. It cautioned that the figure comes from synthetic attacks and "should not be taken as the miss rate on real traffic."

Context: from Claude Code to everyday Claude users

The August change applied to new Claude Code sessions starting Aug. 14. At that time, Enterprise, the Claude API and cloud platforms including Amazon Bedrock and Microsoft Foundry stayed opt-in "for now." Anthropic also stopped charging Pro, Max and Team users for the classifier's overhead.

The app rollout brings the same approach to a broader audience that may not write code. Anthropic has said auto mode "relies on classification systems and therefore does not eliminate risk," and it recommends manual review for high-stakes production infrastructure changes.

What's next

Anthropic has not published a dated rollout schedule or a help-center page for the app change. Users who handle sensitive accounts or irreversible tasks may want to check their permission setting once the switch arrives. Team administrators can watch for whether organization-level controls, already available in Claude Code, extend to the apps.

FAQ

When does auto mode become the default in the Claude apps? Anthropic said "next week" in its Oct. 22 post but did not give an exact date.

Can I turn auto mode off? Yes. Anthropic said users can switch back to manual approval at any time.

Does this apply to Enterprise or free users? The announcement named only Pro, Max and Team plans.

References

• Auto mode is now the default in Claude Code for Pro, Max, and Team plans — Anthropic (Aug. 7, 2026)
• Auto mode for Claude Code — Anthropic (March 24, 2026; updated July 10, 2026)
• Claude Just Got a Major Autonomy Boost: Anthropic Sets Auto Mode as Default — WinCentral (Oct. 23, 2026)