Realtime AI News
Hark launches Hark Handoff, a browser use agent for completing tasks
Hark, the AI startup that raised $700 million in May, has launched Hark Handoff, a browser-use agent that completes tasks like ordering food, booking travel, and shopping even on sites without official APIs. The company claims its action-predicting model makes it faster and cheaper than rivals.
AI startup Hark has launched Hark Handoff, a browser-use agent that takes a command and completes it in the browser — ordering food or coffee, booking travel tickets, filing returns, shopping for essentials, reserving restaurant tables, or researching across multiple sources.
The company raised $700 million in Series A funding in May to build what it calls a universal AI interface. Hark says Handoff can navigate websites that have no official APIs, including Target, Walmart, OpenTable, and LinkedIn.
Rather than relying on vendor-provided interfaces, the agent looks at website structure and visual data to figure out whether it needs to click buttons or type information. Users describe their needs in natural language, including fuzzy terms.
In a video demo, CEO Brett Adcock showed the assistant building a bouquet with flowers the user specifies, accommodating fuzzy terms like some of the florist's choice. TechCrunch notes the video shows only part of the process, so its real effectiveness is hard to gauge.
On the technical side, Hark is using a post-trained model for this release and plans to pre-train later this year. The company says this approach lets it refine its data pipeline, training infrastructure, and techniques more quickly.
Unlike large language models that predict the next token, Hark says its model predicts the next action — a click or a keyboard input at a specific place — which it argues is a more direct and efficient way to drive a browser. The company also claims Handoff is faster and cheaper than the competition.
The computer-use agent field is crowded: Google, OpenAI, and Anthropic are all building computer-use agents, and VC-backed startups like Browser Use are pushing browser automation forward. Whether Hark's predict-the-action-rather-than-the-token approach and cost claims give it an edge will only be proven in real-world use.
Why it matters
Browser-use agents are one of the most competitive fronts in AI agents, and Hark is entering with an action-prediction architecture and cost claims. Its real-world task reliability remains the biggest open question.
Nearby Updates
All08/05, 23:56
Shopify says AI search is driving more traffic and sales, not replacing Google
Shopify President Harley Finkelstein said on the Q2 earnings call that AI has become a complement to search rather than a substitute, with AI-driven traffic and orders to Shopify stores tripling year over year. Revenue rose 36% to $3.6 billion, beating Wall Street's $3.4 billion forecast, and the company credited AI search for part of that strength.
08/05, 23:00
NeuralTrust Launches Runtime Security Mesh for AI Agents, Claiming Coverage of Every Agent
NeuralTrust has announced the launch of what it calls the first runtime security mesh that can extend to every AI agent. The product targets the growing need for runtime protection as agents move from demonstrations into real production workflows.
08/06, 00:35
Cloudflare launches Cloudflare OS, an open-source AI agentic workspace for the enterprise
Cloudflare has launched Cloudflare OS, an open-source AI agentic workspace that runs in the browser and gives employees custom shared micro-apps for research, documents and automated workflows. Built on each company's own Cloudflare account, it is zero-trust by default and supports bring-your-own-model via Cloudflare's AI Gateway.
08/06, 00:35
AI Agent Security Startup Zenity Raises $125M Series C
Zenity, a platform focused on securing AI agents, has raised $125 million in a Series C round to expand its offerings. The raise underscores surging investor appetite for security tools as enterprises deploy AI agents at scale.