Installation

Requirements

• Python 3.12+
• pip, uv, or pipx
• Embeddings: ONNX (built-in) / Ollama (~270MB) / OpenAI API

LTM Server (memtomem)

# uv (recommended)
uv tool install memtomem

# pipx
pipx install memtomem

# pip
pip install memtomem

Optional extras

pip install memtomem[korean]       # kiwipiepy Korean morphological analysis

After installation, the mm CLI is available:

mm init          # interactive setup wizard
mm serve         # start MCP server
mm web           # launch Web UI dashboard

STM Proxy (memtomem-stm)

# uv (recommended)
uv tool install memtomem-stm

# run without installing
uvx memtomem-stm --help

# pip
pip install memtomem-stm

Optional extras

pip install memtomem-stm[langfuse]     # Langfuse observability tracing
pip install memtomem-stm[langchain]    # LangChain agent integration

After installation, the mms CLI is available:

mms add <name> --command <cmd>     # register upstream MCP server
mms list                           # list registered servers
mms serve                          # start STM proxy

Embedding Providers

Provider	Setup	GPU	Cost
ONNX (fastembed)	Built-in	Not required	Free
Ollama	ollama pull nomic-embed-text	Not required	Free
OpenAI	API key required	—	Paid

Choose a provider during mm init or set the MEMTOMEM_EMBEDDING_PROVIDER environment variable.

Tech Stack

Category	Technology
MCP	FastMCP (stdio, SSE, Streamable HTTP)
Framework	Pydantic v2, Click (CLI), FastAPI (Web UI)
Database	SQLite (FTS5 full-text search), sqlite-vec (vector search)
Code parsing	tree-sitter (Python, JS, TS AST)
Korean	kiwipiepy morphological analyzer (optional)
Observability	Langfuse (optional)