OpenClaw releases v2026.4.11 to add ChatGPT import and video generation capabilities

On April 12, OpenClaw released the latest version of v2026.4.11. This update focuses on functional extensions and stability fixes. In terms of core new functions, the Dreaming module has added a ChatGPT dialogue history import function, and added two diary sub-tabs to the UI, Import Insight and Memory Palace. Users can directly browse the imported original dialogue, compiled Wiki pages and complete source pages in the interface to further enhance the long-term memory management experience. In terms of video generation capabilities, this update adds a URL-only generation asset delivery method, typed providerOptions, reference audio input, adaptive aspect ratio support and a higher image input limit, allowing video providers to expose richer generation models while avoiding large files residing in memory. In terms of channels and platforms, Feishu document comment sessions support richer context resolution, comment responses, and input feedback; Microsoft Teams adds response operations, response lists, Graph paging, and delegated OAuth sending support. In addition, the plug-in list can now declare activation and configuration descriptors, simplifying the plug-in access process. In terms of repairs, this time it solved problems such as CodexOAuth's failure to login for new users due to scope rewriting, OpenAI/Groq/Mistral voice transcription was interrupted due to abnormal DNS configuration, macOSTalkMode microphone permissions were granted, and multiple channel stability issues such as confusing WhatsApp default account registration, alternating and misaligned Telegram topic session file paths, and leaking sub-agent internal progress information to the parent session.