The new version of OpenClaw reduces the model list loading delay by 4,000 times and supports Discord voice meeting recordings.

According to Beating Monitoring, the open-source AI assistant project OpenClaw released version v2026.5.22, bringing significant updates to gateway performance and external plugin support. The new version pre-warms the service provider authentication status at gateway startup, drastically reducing the response delay of the model list retrieval interface from about 20 seconds to 5 milliseconds, achieving nearly a 4100-fold performance leap. The pre-warming mechanism allows model listing requests in the hot path to bypass the cumbersome plugin and external command-line discovery processes directly.

In terms of plugin ecosystem, the new version introduces, for the first time, an independent meeting-notes external plugin separate from the core npm package, and supports Discord voice channels as the first real-time input source. Users can access recordings via the read-only command-line tool openclaw meeting-notes or manually import recordings as text. To reduce context overhead during multi-agent deployment and prevent sensitive information leaks, the new version also refactors the startup context mechanism for sub-agents, by default only transmitting AGENTS.md and TOOLS.md to delegated sub-agents, fully isolating personal settings, user memories, and identity key files.

Regarding underlying dependencies and ecosystem integration, the project upgrades the protobufjs dependency to version 8.4.0 to fix security vulnerabilities. For xAI integration, the new version supports reusing Grok’s OAuth authorization files for web searches. Additionally, the OpenAI image generation has been rerouted to the native image interface, thoroughly resolving the historical flaw where legitimate API keys failed authentication in Codex streaming backend, resulting in 401 errors.