Aloha launches AlohaJet to cut AI web automation costs

Aloha unveiled AlohaJet on Sept. 22 in Cyprus, a browser built for AI agents that aims to make web automation faster, cheaper and more reliable. The company says the tool cut token use by 54% and ran 2.2 times faster in a benchmark against standard Chrome automation and Playwright.

Why it matters: - AlohaJet targets one of the biggest bottlenecks in enterprise AI: the rising cost of running agentic workflows across websites and web apps. - The browser is designed to help both individual users and businesses automate web tasks with less compute, lower token usage and better reliability as websites change. - The launch comes as AI spending is already under pressure, with 93% of organizations exceeding budgets and enterprise AI costs rising nearly fourfold.

What happened: - Aloha launched AlohaJet on Sept. 22, 2026, in Limassol, Cyprus. - The product is a browser for AI agents meant to complete complex tasks across websites and web-based applications. - Aloha is the company behind the privacy-focused Aloha Browser, which has served more than 300 million users. - The company positions AlohaJet as a tool for everyday web tasks as well as large-scale business workflows.

The details: - AlohaJet uses LLMdex, a system that helps AI locate the information and actions it needs on a webpage. - Instead of repeatedly interpreting raw page data or screenshots, LLMdex directs models to the relevant parts of a page and identifies the next action. - When multiple models are connected, LLMdex can route routine steps to smaller, lower-cost models and complex steps to more powerful models. - AlohaJet can reuse what it has already learned about recurring workflows instead of relying on fixed instructions that can break when a website changes. - The company says that helps reduce a common reliability problem for web agents when layouts or designs shift. - Completing six common web tasks, including search, form filling, data extraction and multi-page navigation, can require up to 94,000 tokens depending on the automation approach. - Frontier models can cost up to $30 per million output tokens, making token-heavy workflows expensive to run. - In AlohaJet’s benchmark, the same workload used 54% fewer tokens, ran 2.2 times faster and achieved an 88% task success rate. - The benchmark compared AlohaJet with standard Chrome automation and Playwright, a widely used browser automation framework, which posted 51% and 11% success rates respectively. - Users can connect a single AI model or several models, including frontier, open-source or self-hosted options. - For consumers, AlohaJet can search multiple websites for a specific item, compare options and find sellers that deliver to a location. - For enterprises, AlohaJet can run inside a company’s own infrastructure. - When fully configured inside a customer environment, browser sessions, credentials and prompts can remain inside the company’s network.

Between the lines: - The product is aimed at a market where agentic AI can be powerful but expensive, especially when tasks require repeated page reads and step-by-step decisions. - AlohaJet’s pitch is that smarter routing and less page reprocessing can reduce the token burn that makes automation hard to scale. - The emphasis on self-hosting and keeping credentials in-network also suggests Aloha is pitching to buyers with tighter data-control requirements.

What's next: - AlohaJet is intended to be used across individual and enterprise workflows as companies expand AI into more web-based tasks. - The company is betting that lower costs and higher task success will make agentic automation easier to adopt at scale. - If the benchmark performance holds in production, AlohaJet could appeal to teams looking to reduce infrastructure spend while expanding automation volume.

The bottom line: - Aloha is trying to make AI web automation cheaper, faster and more dependable by cutting the amount of model work spent just understanding the page.