Skip to main content
Glama

memsprout for AI agents

Glama

memsprout is a shared AI-context layer for teams — persistent memory your agents search and update over MCP. Every agent and tool you connect reads and writes the same knowledge base, so context survives across sessions, models, and clients. This repo packages the connection two ways: a Claude Code plugin (remote MCP server + a usage skill) and a standalone Agent Skill that works in any client that reads SKILL.md.

Before you install

Create an account at memsprout.com first. The OAuth sign-in that opens when a client first connects does not create accounts — it only signs in existing ones.

Related MCP server: AgentBase

One-command setup

npx memsprout

Detects your installed MCP clients (Claude Code, Codex, Gemini CLI, Cursor), wires each selected one to the server, and offers to scaffold the team files into your repo. Non-destructive, idempotent, stores no credentials. Source in installer/. Not yet published to npm — until it is, use the per-client instructions below.

Claude Code

/plugin marketplace add memsprout/agents
/plugin install memsprout@memsprout

The plugin adds the memsprout MCP server and the using-memsprout skill. The first connection opens the browser OAuth flow.

claude.ai and Claude Desktop

Add memsprout as a custom connector: Settings → Connectors → Add custom connector, with the server URL below. Sign in when the memsprout consent screen opens.

Agent Skill via skills.sh

npx skills add memsprout/agents

Installs the using-memsprout skill, which teaches your agent when and how to use memsprout's tools. You still need the MCP server connected (see below) — the skill describes the doctrine; the server provides the tools.

Cursor

The cursor/ directory has copy-in files for both motions: merge cursor/mcp.json into ~/.cursor/mcp.json for yourself, or commit it as .cursor/mcp.json plus the cursor/rules/memsprout.mdc project rule to your team's repo so every teammate's Cursor shares the same memory. See cursor/README.md.

Codex

codex mcp add memsprout --url https://mcp.memsprout.com/mcp

The Codex CLI, IDE extension, and ChatGPT desktop app share ~/.codex/config.toml, so this one command wires all three. For team setup, append the client-agnostic codex/AGENTS-snippet.md to your repo's AGENTS.md so every AGENTS.md-reading agent follows the team's memory doctrine. See codex/README.md.

Any other MCP client

Point your client at the remote server:

https://mcp.memsprout.com/mcp

Streamable HTTP with OAuth 2.1 — clients that support MCP authorization discover the sign-in flow automatically. Per-client wiring instructions (Codex, Gemini CLI, Cursor, and more) are at memsprout.com/connect.

What's in this repo

.claude-plugin/marketplace.json      Claude Code marketplace manifest
memsprout/                           the plugin
  .claude-plugin/plugin.json         plugin manifest
  .mcp.json                          remote MCP server config
  skills/using-memsprout/SKILL.md    usage skill
mcp-registry/                        official MCP Registry manifest + publish runbook
  server.json
  PUBLISH.md
cursor/                              Cursor setup files (individual + team)
  mcp.json                           MCP config — global or repo-committed
  rules/memsprout.mdc                always-applied project rule (memory doctrine)
codex/                               Codex setup (individual + team)
  AGENTS-snippet.md                  client-agnostic AGENTS.md section (memory doctrine)
installer/                           npx installer (npm package "memsprout")

Support

  • Product: memsprout.com

  • Issues with this plugin or skill: open an issue in this repo

Maintenance

ActivityInactive
ResponsivenessNo issues

Related MCP Connectors

Related MCP Servers

  • A
    license
    Not graded
    quality
    C
    maintenance
    Provides AI agents with persistent knowledge storage, enabling them to store, search, and retrieve text, documents, and files using semantic and keyword search via MCP tools.
    33
    Apache 2.0
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables AI agents to persistently store and semantically search shared knowledge via MCP tools.
    2
    MIT
  • A
    license
    Not graded
    quality
    A
    maintenance
    Governed knowledge base for AI agents via the Model Context Protocol (MCP), enabling agents to search, read, and contribute persisted knowledge with versioning, audit trails, and approval workflows.
    50 npm
    MIT
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables AI agents to capture, share, and recall team knowledge with role-based access control, source attribution, and shared skills over MCP.
    Apache 2.0