June 22, 2026

Dev Tools|Index 02

Pulse: A Local Monitor for Claude Agent Activity

A student-built open-source application offers real-time visibility into Claude agent token usage and costs, with remote tool call approval, all running locally.

Via
AITECH TOKYO Editors
Dateline
TOKYO, 2026-06-20
Date
June 20, 2026
Time
5 min read
Pulse: A Local Monitor for Claude Agent Activity

Tagline

Local monitor for Claude agent activity and costs.

Who & Why

For developers or researchers experimenting with Claude agents who need real-time local visibility into token usage, costs, and remote approval for tool calls, without relying on cloud dashboards.

vs. Existing

Pulse offers a local, open-source alternative to relying solely on Anthropic's cloud-based usage dashboards or custom terminal scripts, providing remote approval functionality not typically found in basic monitoring tools.

Tokyo Take

While useful for individual developers globally, Pulse's niche focus on Claude agent monitoring means it offers little immediate change for most Tokyo professionals, whose workflows rarely involve direct agent development at this granular level.

Pulse is a local application designed to monitor and manage Claude agent activity directly from a user's terminal session.

Developed by a student from Flanders, this open-source tool provides real-time visibility into token usage and associated costs for developers working with Anthropic's Claude models.

Its primary function is to offer a transparent view of an AI agent's operations, a feature often obscured in standard API interactions or cloud-based dashboards. Users can track expenditures as they occur, ensuring better budget control.

A notable feature allows for remote approval of tool calls, enabling developers to oversee and grant permissions to their Claude agents from a mobile device, even when away from their primary workstation.

The application emphasizes local execution and privacy, operating without requiring an account or transmitting data to external cloud services. This local-first approach could appeal to developers or small teams prioritizing data sovereignty and cost transparency when experimenting with or deploying Claude agents.

...easily follow what your claude agent is doing right now in your terminal session with an ambiance design and easily give permissions for your agent.

While Pulse offers a useful utility for individual developers and researchers, its impact on broader Tokyo business workflows remains limited. It addresses a niche technical need for those directly building and debugging AI agents, rather than a widespread professional pain point for general business users.

The Briefing

World AI tech, read from Tokyo. Once a week, in Japanese.

Each Friday: the five global AI tech stories Japanese business professionals should know about this week, translated and read through a Tokyo lens — what it means for Japan, what to act on, what to keep watching.

We respect your inbox. Unsubscribe anytime.