Urgent.News

What's breaking now, across thousands of outlets.

AI

How to share memory between Claude Code, Codex and Cursor with Beacon

You fix a mistake in Claude Code. Tomorrow you open Codex or Cursor and correct the exact same mistake again, because each harness keeps its own memory. Beacon by Asymptote Labs is an open-source, local-first memory layer that sits across the harness layer. It captures session history from Claude Code, Cursor, Codex, OpenCode, Cline and 20+ other harnesses into one OpenTelemetry-based trace…

Beacon by Asymptote Labs is an open-source, local-first memory layer that captures session history from Claude Code, Cursor, Codex, OpenCode, Cline and 20+ other harnesses into one OpenTelemetry-based trace format. It solves the problem that a fix for a mistake in one harness needs to be relearned in another. Beacon records prompts, responses, commands and file edits, allowing users to replay exact agent activity and send activity to a Security Information and Event Management (SIEM) system.

To set up Beacon in Local mode, users can follow a guide that covers installation on macOS, Linux, and Windows. The installation process varies slightly between operating systems. For macOS, the fastest path is to let the coding agent do the installation using a command in Claude Code, Codex or Cursor. For Linux, the installer can be run from the terminal. For Windows, the MSI installer can be downloaded and run.

Once installed, Beacon can be accessed locally through a dashboard or by reading raw event stream directly. It can also recall memory from another agent and forward agent activity to a SIEM. Beacon captures session history in the background, and users can check what it has captured using the `beacon traces` command. Users are advised to curate what Beacon extracts, as the review step is important. The guide also suggests installing lenses for additional functionality, but warns to only install trusted lenses.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at dev.to →

More in AI

How to Use a Remote Ollama or llama.cpp Server From a Desktop AI App on Your Mac in 2026

Your MacBook Air has 16 GB of memory. The PC under your desk, or the Mac Studio in the office, has a big GPU and already runs Ollama or llama-server with a 30B model loaded.

  • Mac users with 16 GB can access remote Ollama/llama.cpp server
  • Off Grid AI Desktop enables connection without transferring files
  • Server must be on same network, ports 11434 (Ollama) or 8080 (llama.cpp)

At the Crossroads of Digital Upheaval: Bitcoin’s Corporate Adoption, The Duality of AI, and The Importance of Data Sovereignty

The digital world is undergoing a profound transformation, driven by a confluence of financial innovation, dazzling technological advancements, and a reevaluation of fundamental digital rights.

  • MicroStrategy adopts Bitcoin as core financial asset, challenging traditional paradigms
  • Sophisticated AI tools, like synthetic voice technology, present dual-edged innovation risks
  • Google adjusts privacy policies, emphasizing data protection debate and user sovereignty

AI Agent or Workflow Automation? A Decision Test for Swiss SMEs

AI Agent or Workflow Automation? A Decision Test for Swiss SMEs Short answer: choose the least autonomous system that can meet the real task.

  • Swiss SMEs face choice between AI agents, workflow automation, or hybrid approach
  • Deterministic automation best for predictable tasks with stable inputs and outputs
  • AI-assisted workflows suitable for uncertain language or classification tasks

More from Sunday 25 October →