2026.3.11MINOR⚠ BREAKINGReleased: 2026-03-11

Multimodal Memory + iOS Redesign + Gemini 3.1 Flash-Lite

/** Memory search now indexes images and audio. iOS gets a fresh home screen. Critical WebSocket security patch. */

#Memory#Multimodal#iOS#Gemini#Ollama#PDF#Security

whats_new.md

✨ What's New

Multimodal Memory Search

Memory searches now powered by gemini-embedding-2-preview with opt-in multimodal image and audio indexing. Your agent can now recall images and audio from past conversations.

iOS Home Screen Redesign

Completely redesigned home screen with live overview of running agents, guided onboarding for new users, and a cleaner docked toolbar.

macOS Model Picker

Switch AI models directly within the chat interface on macOS — no more digging through settings to change your model mid-conversation.

First-Class Ollama Browser Sign-In

Ollama setup now includes browser-based cloud sign-in, making local and cloud-based model management seamless.

Gemini 3.1 Flash-Lite

Support for Google's Gemini 3.1 Flash-Lite model for ultra-fast, cost-effective inference.

PDF Analysis Tool

New first-class PDF analysis tool with native support for Anthropic and Google models.

security_fixes.md

🔒 Security Fixes

🛡Critical cross-site WebSocket hijacking vulnerability patched

🛡VirusTotal integration for scanning all ClawHub skills before installation

🛡Enhanced sandbox isolation for file system operations

breaking_changes.md

⚠️ Breaking Changes

⚠Cron jobs and notifications require running 'openclaw doctor-fix' after update to migrate old settings

upgrade.sh

📦 Upgrade to 2026.3.11

# npm global install

$ npm install -g openclaw@2026.3.11

# migrate cron/notifications

$ openclaw doctor-fix

← Previous:

2026.3.7 — ContextEngine + GPT-5.4

All Releases →