en.Wedoany.com Reported - Tencent Cloud officially released the QClaw v0.2.14 update on April 27. This update focuses on five key areas: expanding the agent framework, switching the underlying model, upgrading remote interaction, expanding the connector ecosystem, and team collaboration. QClaw now officially supports the Hermes framework, allowing users to create and run Hermes-type Agents directly within the platform, with the maximum number of Agents per user increased to 10,000. In the update notes, Tencent Cloud stated that integrating the Hermes framework enables QClaw to support more complex multi-step task orchestration and long-period autonomous operation scenarios, where developers only need to write task descriptions to generate a complete agent workflow.
The upgrade of the underlying model to a free-switching mode is an architectural-level change in this update. Previously, QClaw was fixed to a single model backend. After the update, users can freely choose between multiple models when creating an Agent. Currently, two flagship models are supported initially: DeepSeek-V4 Pro and Tencent Hunyuan Hy 3 preview. DeepSeek-V4 Pro was released and open-sourced by DeepSeek on April 24, supporting a 1 million Token ultra-long context. It achieves the best performance among open-source models in Agent coding evaluations, offering an internal user experience surpassing Claude Sonnet 4.5, with delivery quality close to Claude Opus 4.6 (non-thinking mode). Tencent Hunyuan Hy 3 preview was released in April, positioned as a flagship fast-thinking model, excelling in benchmark tests for mathematical reasoning and code generation. QClaw will be compatible with more SOTA models in the future and will support calling the HITL interface in remote control channels, allowing users to intervene in decision-making and corrections at any time during Agent operation.
The remote control channel WeChat Mini Program for individual users has undergone a major interactive upgrade. New voice interaction capabilities allow users to directly wake up the Agent and execute complex operations via voice commands within the WeChat Mini Program, eliminating the need for manual text input. The file-sharing feature for friends has also been launched simultaneously. Users can directly send text, images, and documents generated by the Agent to WeChat friends or groups within the Mini Program, completing the entire chain from Agent output to social distribution. This upgrade transforms the WeChat Mini Program from a simple remote control terminal into a mobile AI interaction entry point, significantly lowering the barrier for non-technical users.
The connector ecosystem has added four platforms: Baidu Netdisk, Ctrip, Fliggy, and Tencent News. The Baidu Netdisk connector allows the Agent to directly read, search, and manage the user's cloud files, automating document organization, data retrieval, and batch downloads. Connectors for Ctrip and Fliggy cover flight search, hotel booking, and itinerary planning scenarios, enabling the Agent to compare prices across platforms and generate optimal travel plans. The Tencent News connector can retrieve real-time news and extract key information. Combined with the Agent's summarization capabilities, it can quickly produce news briefs. Tencent Cloud disclosed in the update notes that the total number of QClaw connectors has been further expanded with this update, and the external services and data sources accessible to Agents continue to increase. A channel for third-party developers to submit connectors independently will be opened subsequently.
Enterprise-level Agent team collaboration features have initially been implemented in Tencent Docs. Team members can directly create, edit, and manage QClaw Agents within Tencent Docs, collaboratively edit Agent configurations, share Agent templates, and view real-time operation logs. This feature fully integrates the Agent development and operation workflow into the document collaboration system, enabling non-technical roles such as product managers and operations staff to participate in Agent design and iteration. Tencent Cloud also disclosed the future roadmap for QClaw: strengthening Agent orchestration capabilities, supporting multi-agent collaborative task scheduling and state synchronization, and driving the evolution of Agents from single-point tools to a team productivity platform. Since its launch, QClaw has maintained a high-frequency iteration pace and has entered the large-scale testing phase in leading domestic Internet companies.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
