OpenClaw 2026.3.31 Update Adds Built‑In QQ Bot and Visual Task Scheduler

The OpenClaw 2026.3.31 release introduces a native QQ Bot with multi‑account support, visual backend task flow management, enhanced multimodal messaging on LINE, and CJK language optimizations, marking a shift from a simple AI chatbot to an integrated AI entry point for Chinese users.

Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
OpenClaw 2026.3.31 Update Adds Built‑In QQ Bot and Visual Task Scheduler

OpenClaw released version 2026.3.31, adding a built‑in QQ Bot that supports private chat, group chat, guild channels, and multimedia message exchange, extending multimodal capabilities on the LINE platform and introducing a visual backend task‑flow manager together with CJK language optimizations.

The QQ Bot integration offers multi‑account handling, secret reference management, slash commands, reminders, and media I/O, turning the bot into an embedded capability that operates directly within communication scenes rather than requiring users to launch a separate AI application.

Multimodal message support now allows users to trigger tasks using images, video, or audio, so multimodal input serves as a task entry point instead of merely being interpreted as text.

The new visual task‑flow management converts OpenClaw from a purely instant‑response dialogue system into a schedulable, interruptible execution platform with list, detail, and cancellation controls.

CJK‑specific enhancements improve context understanding, stabilize long‑term memory, and deliver more natural text‑to‑speech synthesis for Chinese, Japanese, and Korean.

Nodes/exec: Removed duplicate "nodes.run" shell wrapper; all node shells now execute via "exec host=node" with node‑specific abilities retained for media, location, and notification operations.

Plugin SDK: Deprecated old provider compatibility paths and bundled provider mechanism, issuing migration warnings; only "openclaw/plugin-sdk/*" and local "api.ts/runtime‑api.ts" remain as standard entry points.

Plugin & skill installation: Default "fail‑close" policy blocks installation of dangerous code or failed scans unless explicitly forced with "--dangerously‑force‑unsafe‑install".

Android / notifications: Added forwarding controls with app‑package filtering, silent periods, rate limiting, and safer selector behavior for forwarded events.

Matrix / history: "channels.matrix.historyLimit" adds optional historical context for Matrix groups; introduces water‑mark tracking and safe‑retry snapshot mechanism.

Matrix / network: New "channels.matrix.proxy" setting enables HTTP(S) proxy forwarding for Matrix traffic with account‑level overrides.

Matrix / streaming output: Draft‑style streaming output allows in‑place message updates without sending separate chunks.

Matrix / threads: Added thread‑reply configuration for direct messages, ensuring thread isolation matches the originating room or DM.

Slack / execution approval: Native Slack approval routing and permission control let users approve executions directly in Slack.

CLI / initial configuration: When a remote gateway address is rejected, the CLI now resets the URL to a safe local loopback default.

WhatsApp / message reactions: Agents can add emoji reactions (e.g., ❤️) to incoming WhatsApp messages, enabling more natural interaction.

Microsoft Teams / member info: Integrated Microsoft Graph query provides member details for Teams automation.

Overall, these updates suggest OpenClaw is evolving from a conventional AI tool into an "AI entry point" where natural‑language intent is automatically translated into managed workflows.

task schedulingmultimodalQQ BotOpenClawCJK optimization
Machine Learning Algorithms & Natural Language Processing
Written by

Machine Learning Algorithms & Natural Language Processing

Focused on frontier AI technologies, empowering AI researchers' progress.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.