Vector memory MCP server and Claude Code plugin
npx claudepluginhub aeriondyseti/vector-memory-mcpRAG-powered session memory with waypoints for Claude Code
Semantic memory storage for AI assistants. Store decisions, patterns, and context that persists across sessions.
A local-first MCP server that provides vector-based memory storage. Uses local embeddings and SQLite with sqlite-vec for fast, private semantic search — all in a single file.
There are two ways to install Vector Memory, depending on how much integration you want.
Install as a plugin to get the full experience: MCP server, session lifecycle hooks, waypoint skills, and context monitoring — all managed automatically.
# Add the marketplace
claude plugin marketplace add AerionDyseti/vector-memory-mcp
# Install the plugin
claude plugin install vector-memory@vector-memory-mcp
This clones the repo and runs the MCP server directly from source. Hooks handle session start/clear/compact events, and skills provide /waypoint:set, /waypoint:get, and memory usage guidance.
Install just the MCP server via npm if you want memory storage without hooks or skills, or if you're using a non-Claude Code MCP client.
bun install -g @aeriondyseti/vector-memory-mcp
First install downloads ML models (~90MB). This may take a minute.
Then add to your MCP client config (e.g., ~/.claude/settings.json):
{
"mcpServers": {
"vector-memory": {
"type": "stdio",
"command": "bunx",
"args": ["--bun", "@aeriondyseti/vector-memory-mcp"]
}
}
}
Restart your MCP client after installation. You now have access to:
| Tool | Description |
|---|---|
store_memories | Save memories (accepts array) |
search_memories | Find relevant memories semantically |
get_memories | Retrieve memories by ID (accepts array) |
update_memories | Update existing memories |
delete_memories | Remove memories (accepts array) |
report_memory_usefulness | Vote on whether a memory was useful |
set_waypoint | Save session context for later |
get_waypoint | Restore session context |
index_conversations | Index Claude Code session logs as searchable history |
list_indexed_sessions | Browse indexed conversation sessions |
reindex_session | Force reindex of a specific session |
Store a memory:
You: "Remember that we use Drizzle ORM for database access"
Assistant: [calls store_memories]
Search memories:
You: "What did we decide about the database?"
Assistant: [calls search_memories with relevant query]
Session waypoints:
You: "Save context for next session"
Assistant: [calls set_waypoint with summary, completed items, next steps]
Conversation history (requires --enable-history):
You: "What did we discuss about the API design last week?"
Assistant: [calls search_memories with history_only: true, history_before/after filters]
CLI flags:
| Flag | Alias | Default | Description |
|---|---|---|---|
--db-file <path> | -d | .vector-memory/memories.db | Database location (relative to cwd) |
--port <number> | -p | 3271 | HTTP server port |
--no-http | (HTTP enabled) | Disable HTTP/SSE transport | |
--enable-history | (disabled) | Enable conversation history indexing | |
--history-path | (auto-detect) | Path to session log directory | |
--history-weight | 0.75 | Weight for history results in unified search |
Plugin users: The plugin tracks the repo's default branch. To switch channels, reinstall from a specific branch or tag.
npm users: The stable release is what you get by default:
bun install -g @aeriondyseti/vector-memory-mcp
Pre-release channels are available for testing upcoming changes. These are unstable and may break without notice — use at your own risk.