Jambo has introduced Jarvis, an agentic smartphone that places AI agents at the hardware and operating system layer rather than inside standalone apps. The product was announced online through JamboTechnology on X, with the company framing it as a response to fragmented mobile workflows, crowded app screens, and repeated permission prompts. Its pitch is simple: users state a goal, and the device handles the execution across the system.
Jambo pushes AI agents beyond app boundaries
According to the company, Jarvis emerged from Project Edge, which had been built quietly over several months. Jambo also said it spent five years building a vertically integrated stack and now controls both the hardware and software. That structure allows Jarvis agents to work natively across calls, messages, wallets, browsers, and device tasks, instead of being limited to one application and its permission scope.
Jambo describes this as a shift in mobile computing, arguing that app-based agents run into hard limits around access and control. By owning the device stack, the company says it can remove what it called a “permissions maze.” In its own framing, Jarvis is the first true agentic mobile device, built around system-level execution rather than app switching.
One command can trigger six actions
To show how Jarvis works, Jambo presented a scenario in which a father asks for birthday help in a single request: find a nearby cake shop with good reviews, order a chocolate cake under $15, select a gift based on the child’s age, buy it, and set a pickup reminder for Saturday. Jambo said Jarvis automatically breaks that request into six tasks and completes them in sequence.
The workflow includes searching bakeries, comparing ratings, placing the cake order within budget, browsing gifts by age and interest, completing purchases, confirming delivery, paying through an onchain wallet, and setting a calendar reminder. Jambo argues that this is the kind of out-of-the-box experience needed if AI agents are to reach mainstream users, with the complexity handled underneath the interface rather than by the user.
OpenClaw keeps a strict ban on crypto mentions
At the same time, the wider open-source AI agent space is showing clear governance tensions around crypto. OpenClaw, an AI agent framework led by Peter Steinberger, has collected more than 200,000 GitHub stars. Even so, its Discord server maintains a strict prohibition on crypto-related discussion.
In the source material, a user mentioned bitcoin in a benchmarking context and was blocked by moderators right away. Steinberger later said the rule had been clearly stated when users joined the server, adding that no crypto mention of any kind was allowed. The contrast is sharp: while some projects are building AI agents that interact with wallets and onchain payments, parts of the open-source ecosystem are drawing firm boundaries around crypto discussion.