"Designing a memory system for persona-based agents to improve user experience" matching MCP servers:

Local-first macOS Markdown workspace MCP server. Run persona mcp so Claude Code / Cursor can read and write notes and tasks; Ollama auto-detect.

Persists Claude Code sessions to Notion as team memory, enabling queryable recall of decisions, work state, and project architecture.

Enables users to create and control portable, user-owned memory archives, including saving memories, reflections, values, and stories, exporting them as open JSON, publishing or discovering public archives, and loading consented material as a clearly-labelled memory-based AI persona.

Searches Reddit for answers to questions and saves Q&A sessions to Notion.

Provides curated long-term memory for Codex CLI via MCP, with Markdown-based storage, deduplication, and decay-based forgetting.

Persistent memory management for Claude Code, allowing agents to store and retrieve user preferences, environment notes, and skills across sessions.

MCP server that converts PDF, video, web, and audio inputs into structured Markdown notes with support for checkpointing, batch processing, and Obsidian integration.

A persistent personal memory system for Claude that accumulates and organizes knowledge over time using plain Markdown in a local git repo, providing tools for reading, searching, and capturing short-term memories.

A local MCP server that maintains a persona profile (contact library) for AI assistants, allowing them to create, search, retrieve, supplement, and correct information about people mentioned in conversations via tools like search_profiles, get_profile, create_profile, update_profile, add_facts, update_fact, and delete_fact.

Automatically records AI conversation turns and code changes to local Markdown files to provide persistent context across chat sessions. It enables AI agents to search history through MCP tools and provides a web viewer for browsing past discussions.

Provides AI coding agents with a local, Markdown-first shared memory, enabling them to record decisions, failures, progress, and handoffs and retrieve them across sessions via MCP tools.

Provides cross-session persistent memory for coding agents via MCP tools to store, retrieve, and manage notes with hybrid keyword/semantic search and automatic deduplication.

Provides a shared, persistent memory layer for AI assistants, letting them read and write to a local Obsidian vault through MCP with validation, secret rejection, and deduplication.

An MCP server that lets users capture quick phone thoughts via Android and later query them from AI assistants like Grok or Claude, functioning as a simple task memory store without notifications or geolocation.

A local shared-memory MCP server for Claude, Codex, Hermes, and other MCP-aware agents, providing a durable memory layer with Markdown files and SQLite fact store.

A local MCP server that provides agents with tools to list, read, search, inspect history and diffs, and capture unstructured text in a user-owned Git repository of durable memory.

Local-first agentic knowledge layer over Obsidian notes, enabling MCP-aware agents to search, retrieve, and compile knowledge with provenance and task contracts.

Personal memory MCP server for Claude Code with an Obsidian-style graph. Enables persistent, queryable memories across sessions using Markdown files with bidirectional wiki-links and a D3.js force-directed graph UI.

Enables AI assistants to maintain persistent local memory across sessions by organizing experiences into hierarchical layers (raw logs, checkpoints, daily cards, and stories), with tools to write, search, and verify entries via hashes.

Provides tools for AI agents to manage long-term memories, daily notes, and TODO lists through a structured markdown file system. It enables context awareness by allowing agents to read, write, and search entries for persistent information storage.