Skip to main content
AdOpenFree logoPromote your productReach more potential users and drive product growth and revenue.Advertise
Favicon of Linger

Linger

Free Listing

Linger is an open-source personal agent that runs fully on your PC, bringing powerful local AI to everyday computers through efficient model–harness co-design so everyone can run their own.

Visit simplexai-labs/linger

Overview

Linger is an open-source personal agent that runs fully on a user's PC. It combines local model inference, persistent wiki-style memory, proactive assistance, desktop and browser use, and agent teams so everyday Macs and PCs can run their own AI.

Key Features

  • Local model inference through PolyStrata keeps a model's busiest experts on the GPU and the rest in RAM or on SSD; the 125B Qwen3.8-Flash-Next example runs on a 12 GB card with 64 GB of RAM.
  • Model Autopilot detects GPU, VRAM, RAM and bandwidth, recommends the smartest model that fits with at least 64K context, installs it, and counts it as installed only after it answers a real question.
  • Wiki and memory files mail, chats and meetings into wiki pages each night for every agent to search, and the project reports 95% on the LoCoMo memory benchmark.
  • Proactive Another Me watches inbox, calendar and Slack, follows up on work, speaks up only when something needs the user, and rewrites its decision rules nightly through Dreaming.
  • Desktop and browser use lets agents open apps on Mac, Windows or Linux and click and type; on Mac and Windows they work in the user's everyday Chrome with existing sign-ins.
  • Agent teams, background task delegation and model scheduling coordinate multiple agents and long-running work while keeping interactive chat responsive.

Use Cases

  • Run a personal AI assistant fully on a consumer Mac or PC without cloud token fees for local inference.
  • Turn email, chat, calendar, meeting and file context into searchable personal wiki memory shared by agents.
  • Use desktop and browser agents to operate everyday apps and Chrome with the user's existing sign-ins.
  • Coordinate multi-agent work and long background jobs, and record meetings with live transcripts and follow-ups.

Getting Started

  • Install the desktop app for macOS 14.0 or later on Apple silicon, Windows x64, or Linux x64 with glibc 2.35 or newer, and check downloads against SHA256SUMS.txt from the same release.
  • On the desktop app setup screen, use Quick Install to set up a Linger server on the same computer.
  • For a shared server or source build, use Docker Compose: clone the repository, copy .env.example to .env, run docker compose up --build --detach --wait, create an owner account, open the local app on port 3000, and sign in.
  • On the connected computer, run the guided installer for macOS/Linux or the PowerShell installer for Windows; it checks hardware, installs and verifies a local model after approval, and connects linger-computer to the Linger server.

Deployment & Requirements

  • Desktop packages are provided for macOS 14.0 or later on Apple silicon, Windows x64, and Linux x64 with glibc 2.35 or newer; there is no package for Intel Macs.
  • Docker Compose deployment requires Docker with Compose and at least 5 GiB of memory for its virtual machine.
  • The Compose containers run only the Linger server; models and agents run on the computer connected to it through the guided installer.
  • Automatic port override for busy Compose ports requires Docker Compose 2.24.4 or later.

Before You Adopt

  • License: Other. Review its terms before using, modifying, or distributing the project.
  • Intel Mac users must use Docker Compose because no desktop package is provided for that platform.
  • The Compose containers do not run local models or agents; those run on the connected computer.
  • Linger depends on several third-party open-source projects for core capabilities.
  • Memory is kept on the user's device rather than in a provider account.

Comments

Sign in to leave a comment.

More like Linger

Favicon of Mercury Agent

Mercury Agent

Free ListingStars: 3.8K

Soul-driven AI agent that asks before it acts.

Agent Memory

Mercury is an open-source, soul-driven AI agent that runs on your machine, learns how you work, remembers what matters, asks before it acts, works with any model, and is reachable from terminal, web, Telegram, Discord, Slack, or Signal.

Favicon of Hindsight

Hindsight

Free ListingStars: 53.2K

Agent memory that learns

Agent Memory

Hindsight is an agent memory system that helps AI agents learn over time through typed facts, consolidated observations, multi-strategy recall, and reflective reasoning; Memlake serves 100M memories at under 100ms p99 on object storage.

Favicon of Beacon

Beacon

Free ListingStars: 2.5K

Self-improving memory for coding agents

Agent Memory

Beacon is a self-improving, cross-harness memory layer for coding agents. It turns sessions across Claude Code, Codex, Cursor, OpenCode, and 20+ other harnesses into a compounding knowledge layer, so each agent starts smarter.