Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Kimi launches a browser extension that turns web actions into reusable Skills

Moonshot's Kimi has shipped a browser extension that upgrades its earlier Kimi WebBridge: users can now chat with Kimi from a browser sidebar and let it operate the current page, and record actions they perform on a site into a reusable Skill. The older path, where a local agent drives the extension, is still supported.

Published

Kimi now has a browser extension. According to a September 25 report from QbitAI, Moonshot has rebuilt the Kimi WebBridge it launched a few months ago into a "Kimi browser extension," and the biggest change is the entry point: instead of only being driven by a local agent, the extension can now be used directly from the browser sidebar.

Clicking the Kimi icon in the browser toolbar and signing in lets a user talk to Kimi in the sidebar and have it operate the current page, while the original route — a local agent calling the extension — remains available. Opening pages, reading content and clicking buttons all carry over from WebBridge.

Two capabilities are genuinely new. One is the sidebar itself. The other is recording: the extension can capture what a user does on a page and turn those steps into a Skill. A daily routine — open the same site, look up the day's news, export the data — can now be stored as a Skill and called again later, and if a session goes particularly well, Kimi can save that conversation as a Skill too.

The company frames this as making the extension better suited to repetitive work: available at hand, and repeatable.

The report is candid about the limits. Official guidance notes that site redesigns or dynamically loaded pages can still break a run midway, and suggests taking a screenshot to confirm page state before adjusting instructions when a complex page fails.

Seen against the short history of AI browser tools, the shift is a familiar one. An earlier generation — Monica, Doubao's plugin — was mostly about "reading the web": asking questions, translating pages, skimming papers. Manus's Browser Operator and Kimi WebBridge moved to "using the web," letting an agent drive the user's real browser with permission. Turning past workflows into Skills pushes one step further, toward accumulation.

Timing is part of the story. The extension lands right on the heels of Kimi Code Desktop, a workbench for agent projects whose built-in browser follows the project so changes can be checked immediately. One product points inward at the project; the other points outward at the Chrome or Edge the user already lives in.

What to watch next: whether recorded Skills stay reliable across sessions and how often they fail on frequently redesigned sites — and whether the extension gets wired more tightly into the Kimi Code Desktop workflow, so internal debugging and external browsing share one set of capabilities.

Why it matters

Letting Kimi work from the browser sidebar puts its agent inside the Chrome or Edge users already have, and Skill recording could turn repeated web chores from per-task coaching into record-once, reuse-many.

KimiAgent
Back to realtime news

Nearby Updates

All

09/25, 14:14

Ex-Huawei large-model duo's Physical AI startup Xirang Kaiwu raises hundreds of millions of yuan

Physical AI startup Xirang Kaiwu, founded by two former Huawei large-model leaders, has closed several hundred million yuan in consecutive seed and angel rounds at a USD 500 million valuation. The company is pretraining an LPM — a Large Physics Model — on video, robot trajectories and real interaction data, and reports early leaderboard results for its model RiXin.

09/25, 13:52

Mifeng, billed as 'China's Index', recruits ordinary people to teach robots for extra income

A Chinese robot-data platform called Mifeng has arrived, described by media as 'China's Index' and pitched at ordinary people who teach robots while earning money on the side. It targets the scarcest input in embodied AI, real operation data, though whether crowdsourcing can deliver usable quality is still an open question.

09/25, 13:00

Lightspeed targets $250M for a new India fund focused on early-stage AI

Lightspeed is targeting $250 million for a new India fund that will focus on early-stage AI, TechCrunch reported on September 25. The Silicon Valley firm is aligning its India fundraising cycle with its global funds for the first time and moving to a shorter investment period.

09/25, 06:40

Alibaba Cloud’s Bailian platform launches a preview decision model with Token Plan support

According to 53AI, Alibaba Cloud has added a preview version of a decision model to its Bailian platform, along with support for a Token Plan. The preview is free for a limited time, giving developers a low-cost way to test model-driven decision-making inside business workflows.