Contextberg

Local AI agent memory for macOS & Windows, served via MCP

160 followers

Local AI agent memory for macOS & Windows, served via MCP

160 followers

•

•

Contextberg is a local AI agent memory app for macOS and Windows. It captures screens, browser history, and agent conversations into a private, searchable archive, then serves relevant context to Codex, Claude Code, Cursor, and other agents via MCP. Search screenshots with OCR, exclude any app or source, and use your existing Codex sign-in, Gemini/OpenRouter API key, Contextberg Cloud, or a local model. Your archive stays on-device; only selected context goes to your chosen provider.

This is the 2nd launch from Contextberg. View more

Contextberg

Launching today

Turn your work into private memory for AI agents

Contextberg brings local AI agent memory to macOS and Windows. It captures screens, browser history, and agent conversations into a private, searchable archive, then serves relevant context to Codex, Claude Code, Cursor, and other agents over MCP. This launch adds native macOS capture, OCR screenshot search, source exclusions, and flexible model routing: use your existing Codex sign-in, a Gemini/OpenRouter API key, Contextberg Cloud, or a fully local model.

Free

Launch tags:Productivity•Developer Tools•Artificial Intelligence

Launch Team

Contextberg

Maker

📌

Hey Product Hunt 👋 I’m Tiger, the solo founder of Contextberg. I built it because I was tired of re-explaining my work to AI agents. After every reset, task switch, or weekend away, the context already existed—in my screens, browser research, and previous agent conversations—but I had to reconstruct it manually. Contextberg turns that work into local, reusable memory and serves the relevant context to Codex, Claude Code, Cursor, and other agents over MCP. What’s new in this launch: 🍎 Native macOS app, alongside Windows 🔎 OCR search across captured screens 🛡️ App and source exclusions 🧠 Short-term, daily, and long-term memory 🔀 Use your Codex sign-in, Gemini/OpenRouter key, Contextberg Cloud, or a local model Your archive stays on-device. Only the context you choose to use is sent to the model provider you select. I’m building the memory layer under the agent—not another agent that locks you into one model. What part of your workflow does your AI agent forget most often? I’d love to hear how you currently reconstruct context.

Report

1mo ago

Previous Contextberg Launches

ContextbergTurn your work into AI agent memory, served over MCP

Launched on May 20th, 2026