
According to Tencent Cloud’s official announcement on April 27, the Tencent Cloud AI Agent desktop tool QClaw has officially released version v0.2.14. Tencent Cloud stated that this is QClaw’s largest-scale version upgrade to date. Key updates include integration with the Hermes Agent framework, an upgrade of the underlying models to a freely switchable mode, and a comprehensive upgrade of the “Inspiration Plaza” to the “Expert Plaza.”
According to Tencent Cloud’s official announcement, QClaw v0.2.14 has officially integrated the Hermes Agent framework. Users can create and run two Agents with different cores within the same application. At present, Hermes-type Agents only support macOS.
Underlying model free switching: Latest models such as DeepSeek-V4 Pro added
According to Tencent Cloud’s announcement, in QClaw v0.2.14 the underlying model has been upgraded from a fixed usage mode to a freely switchable mode. Users can choose system intelligent matching or manually specify a model. Models already supported include DeepSeek-V4 Pro, Hy 3 preview, Kimi-K2.6, and GLM-5.1.
According to Tencent Cloud’s announcement, the original “Inspiration Plaza” has been comprehensively upgraded to the “Expert Plaza.” In the first phase, more than 100 prebuilt Agents categorized by industry and scenario have been launched, covering fields such as content creation, data analysis, and code development. Tencent Cloud explained that the usage process has three steps: choose an expert, describe your needs, and get results. Each prebuilt Agent comes with an independent persona, an isolated conversation space, and multi-step workflows. Users do not need to write a Prompt or make any configurations.
Other updates include: WeChat mini-program support for voice remote control and document sharing; four new connectors—Baidu Netdisk, Ctrip, Feishu, and Tencent News; and the Agent team collaboration feature based on Tencent Docs being同步 launched.
According to Tencent Cloud’s official announcement, Hermes-type Agents currently only support macOS. Users can simultaneously create and run two Agents with different cores within the same QClaw application.
According to Tencent Cloud’s announcement, the underlying models currently supported include DeepSeek-V4 Pro, Hy 3 preview, Kimi-K2.6, and GLM-5.1. Users can choose system intelligent matching or manually specify a model.
According to Tencent Cloud’s announcement, the “Expert Plaza” upgrades the underlying capabilities from Skills to a full Agent architecture. Each prebuilt Agent has an independent persona, an isolated conversation space, and a multi-step workflow. In the first phase, more than 100 prebuilt Agents have been launched. Users can use them directly without having to write a Prompt.
Related News
xAI Grok Voice takes over Starlink customer service hotline, with 70% of calls automatically closed
Telegram launches a no-code AI bot tool, supporting GPT and Llama models
DeepSeek V4-Flash goes live on Ollama Cloud, US-hosted: Claude Code, OpenClaw one-click integration
OpenAI launches GPT-5.5: 12M context, AA index tops the chart, and Terminal-Bench rewrites the agent benchmark with 82.7%
Google Jules releases a new version candidate list, repositioning it as an end-to-end product development platform