Releases·Tool Release
Developer open-sources OpenContext, a zero-cloud markdown context store for AI agents via MCP
A developer released OpenContext, an MIT-licensed MCP server that stores agent memory as plain markdown files in a project's .opencontext/ directory instead of third-party vector databases or cloud sync. It runs over strict stdio transport with no open ports or background daemons, keeping memory human-readable and Git-friendly. It offers a lightweight, self-hosted alternative for developers wanting transparent agent context management.
2 engagement1 sourceSat, Aug 29, 2026, 11:32 PM
AI Overview
- Stores all agent memory as plain .md files in a local .opencontext/ working-directory folder, avoiding third-party vector databases and proprietary cloud sync
- Implements the Model Context Protocol with strict stdio transport, so no open ports or background daemons run
- Provides transparent file operations: read, save, and list topics
- Zero dependencies beyond the official MCP TypeScript SDK
- MIT-licensed; code on GitHub, site at opencntx.dev
Related
Products·launched
Local-first, git-backed memory for Claude Code
An open-source tool that gives Claude Code persistent memory using local markdown files versioned with git. It auto-injects relevant context at the start of each session via hooks, eliminating the need to re-explain project details. Works with any MCP agent, no cloud or API keys required.
An open-source tool that gives Claude Code persistent memory using local markdown files versioned with git. It auto-injects relevant context at the start of each session via hooks, eliminating the need to re-explain project details. Works with any MCP agent, no cloud or API keys required.
8 engagement1 sourceredditFri, Jul 24, 2026, 08:21 AM
Products·launched
samemind: AI agent memory as plain markdown in git, no cloud or API keys
samemind stores AI agent memory as plain markdown files in a git repository, eliminating the need for databases, cloud services, or API keys. It works across 12 AI coding engines including Claude Code, Cursor, and opencode, providing a portable and version-controlled memory layer for AI agents.
samemind stores AI agent memory as plain markdown files in a git repository, eliminating the need for databases, cloud services, or API keys. It works across 12 AI coding engines including Claude Code, Cursor, and opencode, providing a portable and version-controlled memory layer for AI agents.
1 engagement1 sourceredditTue, Jul 14, 2026, 06:16 PM
Products·launched
CodeLore: MCP server for persistent agent memory via markdown docs
CodeLore is a small MCP server that gives LLM agents a persistent place to store and retrieve knowledge about a codebase. It writes markdown files in a fixed hierarchy (project > category > chapter > topic) with internal and usage docs, so the agent doesn't need to re-read the entire codebase every session. It solves the problem of wasted context and repetitive learning for developers using AI coding assistants.
CodeLore is a small MCP server that gives LLM agents a persistent place to store and retrieve knowledge about a codebase. It writes markdown files in a fixed hierarchy (project > category > chapter > topic) with internal and usage docs, so the agent doesn't need to re-read the entire codebase every session. It solves the problem of wasted context and repetitive learning for developers using AI coding assistants.
6 engagement1 sourceredditFri, Jul 17, 2026, 09:37 AM
Products·launched
Synapse - local tool that turns ChatGPT exports or text folders into a Markdown archive with SQLite search
Synapse is a rebuild of a LangGraph hackathon-winning agent memory and personal wiki system, made runnable locally without servers or databases. It ingests a ChatGPT export or messy text folder into local Markdown plus a SQLite search index, needing no API key. An optional OpenAI-compatible model can convert raw history into linked notes.
Synapse is a rebuild of a LangGraph hackathon-winning agent memory and personal wiki system, made runnable locally without servers or databases. It ingests a ChatGPT export or messy text folder into local Markdown plus a SQLite search index, needing no API key. An optional OpenAI-compatible model can convert raw history into linked notes.
4 engagement1 sourceredditWed, Sep 9, 2026, 11:07 AM
Releases·Tool Release
ContextMemory: open-source proxy adds session memory and shared wiki to any Ollama-compatible client
A developer released ContextMemory, an open-source proxy that sits between any Ollama-compatible chat client and the LLM, adding per-session markdown memory and optional tools without changing the client code. It works by swapping the chat URL to point at ContextMemory instead of raw Ollama.
A developer released ContextMemory, an open-source proxy that sits between any Ollama-compatible chat client and the LLM, adding per-session markdown memory and optional tools without changing the client code. It works by swapping the chat URL to point at ContextMemory instead of raw Ollama.
2 engagement2 sourcesredditSun, Jul 26, 2026, 11:44 PM