OpenClaw releases v2026.4.11, adding ChatGPT import and video generation capability extensions

OpenClaw releases the latest version v2026.4.11, focusing this update on feature extensions and stability fixes.
Regarding core new features, the Dreaming module adds ChatGPT conversation history import functionality, and two new diary sub-tags, "Import Insights" and "Memory Palace," are introduced in the UI, allowing users to browse imported raw conversations, compiled Wiki pages, and full source pages directly within the interface, further enhancing long-term memory management experience.
In terms of video generation capabilities, this update introduces URL-only asset delivery, typed providerOptions, reference audio input, adaptive aspect ratio support, and higher image input limits. Video providers can leverage these to expose richer generation modes while avoiding large files residing in memory.
On channels and platforms, Feishu document comment conversations now support more comprehensive context parsing, comment reactions, and input feedback; Microsoft Teams adds reaction operations, reaction lists, Graph pagination, and delegated OAuth sending support. Additionally, plugin manifests can now declare activation and configuration descriptors, simplifying plugin integration.
Fixes include resolving new user login failures caused by scope rewriting in Codex OAuth, interruptions in OpenAI/Groq/Mistral speech transcription due to DNS configuration issues, issues with second-time toggling of microphone permissions in macOS Talk Mode, confusion caused by default WhatsApp account registration, misaligned Telegram topic session file paths, and internal progress information leakage from sub-proxy to parent sessions, among other channel stability issues.