MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
Embedded NoSQL Database for TypeScript & AI Agents. Zero-native document database with ACID transactions & in-memory cache — verified on Node.js & Bun.
npm MIT -
SynapCores Community Edition: free self-hosted AI database (SQL, vector search, graph, in-database ML, agent memory, MCP). Install: curl -fsSL https://get.synapcores.com | sh
★ 1 -
Connect ChatGPT to a free MongoDB database in the cloud via MCP — persistent memory across chats
MIT -
Cognitive memory database for AI agents — consolidates duplicates, detects contradictions, and fades stale memories via temporal decay, written in Rust (Apache-2.0) as a library, MCP server, or HTTP cluster.
cargo Apache-2.0 ★ 180 -
Managed MCP server for the Brain memory database — Brain's full tenant surface as 25 MCP tools over remote Streamable HTTP, in front of the hosted gateway or a self-hosted brain-edge.
npm Apache-2.0 ★ 1 -
Persistent memory database for MCP systems — SQLite, FTS5 search, vector embeddings, 3-signal RRF recall, ACT-R decay, graph edges. (Companion repo — architecture and interface only. Source lives in the private engine.)
-
Build a second brain every AI can share: your content in Oracle AI Database 26ai — semantic search with in-database embeddings, a self-compiling wiki, four kinds of agent memory, and an MCP server that plugs it into Claude, ChatGPT, and any MCP client. Runs on the free local container or Always Free cloud.
MIT ★ 13 -
pg-mnemosyne-mcpgithub.com/Janadasroor/pg-mnemosyne-mcpupdated this year
🧠 A high-performance PostgreSQL-backed MCP server acting as a super memory, task tracker, and dynamic database manager for AI agents. Features built-in connection pooling, a professional tasks schema, and a unique shared multi-agent coordination hub to prevent coding conflicts in real-time.
MIT ★ 4 -
Native Zero-Dependency In-Memory Database Anonymizer & GDPR Test Data Generator (CLI & MCP Server)
pypi MIT -
In-memory graph database and Gremlin traversal engine in Rust: browser playground, CLI, MCP server for AI agents, durable store, Python bindings.
cargo Apache-2.0 ★ 23 -
persistorgithub.com/briancolinger/persistorupdated this year
Multi-tenant remote MCP memory server in Go. OIDC to Postgres row-level security; tenancy enforced by the database.
NOASSERTION -
A persistent, Git-versioned map of your whole codebase and database schema that coding agents read before they touch anything. Local-first MCP server + CLI in Go: a governed repository index of code knowledge that gives Claude Code, Codex, Cursor, and opencode long-term context, memory, and code intelligence.
go NOASSERTION ★ 1,923 -
Markdown memory for AI agents. Files you can read, edit, grep, and commit. Not a database. Listed in the MCP Registry.
npm MIT ★ 1 -
Runs SQL on real PostgreSQL and SQLite in memory: your schema, the database's own errors, results.
MIT ★ 1 -
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.
go MIT ★ 865 -
Long-term memory for AI agents as plain JSON objects in S3 (or R2, SeaweedFS, MinIO, a directory): no database, BM25 + local embeddings, conditional-write concurrency, MCP server, LangGraph store.
MIT -
Persistent semantic memory for Claude Cowork. Extracts beliefs, insights, and research from every session into a local database Claude queries in real time. Nightly research scouting, automated belief verification, and MCP-based retrieval. Built for Cowork — not Claude Code.
MIT -
This MCP Server keeps learnings from the development as long time memory in a notion database.
-
KR Mycelia Memory – A native Windows 11 AI memory system powered by a Dynamic Associative Database (DAD). Persistent local memory, offline-first AI integration, MCP agent access, semantic associations, document and media processing, and transparent source-backed recall. Built in C++20.
NOASSERTION -
Agent memory in a single file — token-budget recall, MCP server, agent skill. No server, no database.
MIT ★ 1 -
Passthrough MCP server that materializes tool results into an in-memory DuckDB scratch database.
pypi Apache-2.0 -
Persistent memory & issue tracking for AI coding agents — todos, changelogs, plans, sprints & session handoffs in one cross-project database over MCP. Built for Claude Code.
pypi MIT -
Local-first database for AI agents. Runs, retries, forks, traces and memory in one SQLite file. Python + TypeScript SDKs, CLI, and MCP server. Zero dependencies.
Apache-2.0 -
mcp-context-graphgithub.com/padobrik/mcp-context-graphupdated this year
A zero-dependency, in-memory graph database providing a memory layer for local AI agents with AST-based dependency tracking and Token-Level Provenance via Source Maps.
pypi MIT -
Agent memory, made simple. Plain Markdown in your own Git repo, shared by all your coding agents, like Claude Code and Codex. One command to install. No API key, no database, no server to run.
npm MIT ★ 4 -
Define AI tools in YAML with natural language schemas. All tool usage is automatically stored in Qdrant vector database, enabling semantic search, filtering, and memory retrieval across sessions.
MIT ★ 24 -
Agent memory you can actually read — it lives in your own markdown, not a vendor's database. Retrieval over the notes you already wrote, and a credential vault agents can use but never see. Self-hosted, MCP, one Go binary. Works on any folder: Obsidian vault, Logseq graph, or plain ~/notes.
go MIT ★ 3 -
Durable project memory for AI coding agents (Claude Code, Codex, Cursor, Gemini CLI): decisions and bug post-mortems as markdown pages in your repo, BM25F search via CLI or MCP server, validated writes. No server, no database, stdlib-only Python.
pypi NOASSERTION -
A Model Context Protocol (MCP) server that loads ROS2 rosbag files containing log information (/rosout), builds an in-memory database, and provides SQL-like querying capabilities.
MIT ★ 4 -
Durable memory for AI investigations: evidence, claims, hypotheses, questions and conclusions in a local database shared by a desktop app, a CLI and an MCP server.
MIT -
PhotonDB is a modern C++ rendition of a high-performance in-memory key-value database with MCP server. The data model is key-value, active development is underway to support a wide variety of value types
-
A Model Context Protocol (MCP) server that gives LLMs persistent project memory via SQLite. This eliminates the need to re-explain project context in every new conversation. The LLM queries the database and instantly knows where things stand.
npm Apache-2.0 ★ 9 -
The Dry.ai Model Context Protocol (MCP) server enables seamless integration between Claude and your Dry smartspaces, turning conversations into persistent, structured data. Currently, you can use this server to create and store memories and query the information in your database (smartspace).
npm MIT -
ContextLib: a project's long-term memory as plain Markdown files you can read in Finder, on an SSD, with no database required. Append-only decisions, facts and lessons, a tamper-evident ledger, an 8 KB always-loaded brief, a CLI and an MCP server. Python stdlib only.
oci Apache-2.0 ★ 1 -
Selvedgeno update data
**Persistent decision memory for AI coding agents.** Selvedge records why code changed, which approaches were rejected, and when decisions deserve another look, so future sessions can retrieve that context before editing. Query a function, database column, API route or dependency instead of reconstructing its history from an old conversation. Claude Code, Codex, Cursor, Copilot, Gemini CLI and Windsurf can connect through MCP. The CLI works independently. Decisions live in a local SQLite store, with no required hosted account and no LLM in the core storage or retrieval path. A saved explanation is testimony supplied by a person or agent. Selvedge does not extract hidden model reasoning, verify that an explanation is true, or guarantee that an agent will consult it. [Compatibility and capabilities](https://selvedge.sh/reference/compatibility/) · [56-second technical demo](https://youtu.be/wXoA9htThAM)
smithery -
Git for agent memory, shared by every AI tool on your machine: fork, merge, rewind and hand off work between agents over MCP.
cargo Apache-2.0 ★ 1 -
cozo-memorygithub.com/tobs-code/cozo-memoryupdated this year
Persistent, local-first memory for AI agents. No cloud, no Docker, no external services – just CozoDB embedded in Node.js.
npm Apache-2.0 ★ 1 -
In-memory data and messaging server for agents, on port 2790: data structures, streams, pub/sub, and a forum for agents over HTTP, SSE, and MCP
npm MIT -
agidbgithub.com/rohansx/agidbupdated this year
A cognitive substrate for AI agents — hyperdimensional memory with first-class goals, beliefs, sensory, self-model, and a non-destructive unlearn primitive. Rust top to bottom, embedded-first, no LLM in the read path. One binary, no query language.
Apache-2.0 ★ 6 -
App-first AI framework with a built-in, governed memory engine (memory, agents, RAG, MCP)
MIT ★ 4 -
mcp-memory-sqlitegithub.com/Daichi-Kudo/mcp-memory-sqliteupdated this year
SQLite-based MCP memory server with WAL mode for concurrent access. Drop-in replacement for @modelcontextprotocol/server-memory
npm MIT ★ 4 -
A memory system using mem0 for AI applications. Enables long-term memory for AI agents as a drop-in MCP server.
npm MIT ★ 97 -
A lightweight Model Context Protocol server that gives LLMs persistent, searchable memory. Stores entries in Postgres with pgvector for semantic search, embeds via a local Ollama model, and speaks MCP's Streamable HTTP transport.
MIT -
mcp-bitemporal-memorygithub.com/NotEl1gible/mcp-bitemporal-memoryupdated this year
An MCP server written from scratch in pure stdlib, with two clocks: valid time and transaction time. Answers what the system BELIEVED at a past date, not just what was true.
MIT -
RAMen is a fast in-memory data store like Redis, but built for AI: drop-in Redis protocol, native vector search, semantic caching, and a built-in MCP server for agents. Single Go binary, BSD-3.
go BSD-3-Clause ★ 38 -
Persistent project memory for coding agents. Corrections supersede stale facts in a FalkorDB graph, flag dependent plans, and hand them to agents over MCP.
-
Truly Zero-Dependency Advanced MCP server with local file-based memory features for VS Code/Cursor - NO EXTERNAL API REQUIRED - Built-in AI embeddings using native algorithms - Zero-config local semantic search (Embedding v1.2.0, FileSystem v2.0.0, Memory
npm MIT -
AI Intelligence Platform -- knowledge graph + semantic search + reasoning + multi-agent debate in a single binary
NOASSERTION ★ 20 -
The giant database of institutional knowledge hidden in your code
MIT ★ 3 -
VeritasGraph — open-source Knowledge Graph & GraphRAG framework on GitHub. Build multi-hop reasoning, ontology-aware retrieval, and verifiable attribution over your own data. Nodes, edges, RDF, linked-data — runs locally or in the cloud.
pypi MIT ★ 327 -
Open Brain — The infrastructure layer for your thinking. One database, one AI gateway, one chat channel — any AI plugs in. No middleware, no SaaS.
NOASSERTION ★ 4,776 -
local_faiss_mcpgithub.com/nonatofabio/local_faiss_mcpupdated this year
Local FAISS vector store as an MCP server – Agent Memory, drop-in local semantic search for Claude / Copilot / Agents.
pypi MIT ★ 36 -
Mimirgithub.com/orneryd/Mimirupdated this year
Mimir - Fully open and customizable memory bank with semantic vector search capabilities for locally indexed files (Code Intelligence) and stored memories that are shared across sessions and chat contexts allowing worker agent to learn from errors in past runs. Includes Drag and Drop multi-agent orchestration
NOASSERTION ★ 290 -
MCP local - Code intelligence tools — with broad commands that allow agents in a huge repository to do almost anything: identify the right problem → find the right file/symbol → verify the right flow → measure the scope of impact → safely refactor → check API contracts → audit health, all quickly and accurately.
MIT ★ 12 -
Provenance-tracked shared memory for AI agent teams. A deterministic Guard runs multi-hop graph queries in FalkorDB to block injected actions, and a Warden agent investigates incidents via MCP tools.
★ 2 -
Self-hosted knowledge service for AI agents: an LLM-maintained wiki and a business graph, searched in milliseconds, far faster than RAG, with no LLM or database on the read path. HTTP and MCP.
Apache-2.0 ★ 6