Memforge Memory runs on your own machine — the client starts it, so there is no endpoint to ping. 20 installs a week from npm.
Team engineering memory for AI coding assistants. 47 MCP tools, pgvector semantic search.
Nothing serious here today
Today is the operative word: we check Memforge Memory every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Connect this server
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add memforge-memory -e 'DATABASE_URL=<DATABASE_URL>' -e 'OPENAI_BASE_URL=<OPENAI_BASE_URL>' -e 'OPENAI_API_KEY=<OPENAI_API_KEY>' -e 'OPENAI_EMBEDDING_MODEL=<OPENAI_EMBEDDING_MODEL>' -- npx -y @memforgeai/memory-service{
"mcpServers": {
"memforge-memory": {
"command": "npx",
"args": [
"-y",
"@memforgeai/memory-service"
],
"env": {
"DATABASE_URL": "<DATABASE_URL>",
"OPENAI_API_KEY": "<OPENAI_API_KEY>",
"OPENAI_BASE_URL": "<OPENAI_BASE_URL>",
"OPENAI_EMBEDDING_MODEL": "<OPENAI_EMBEDDING_MODEL>"
}
}
}
}[mcp_servers.memforge-memory]
command = "npx"
args = ["-y", "@memforgeai/memory-service"]
env = { DATABASE_URL = "<DATABASE_URL>", OPENAI_BASE_URL = "<OPENAI_BASE_URL>", OPENAI_API_KEY = "<OPENAI_API_KEY>", OPENAI_EMBEDDING_MODEL = "<OPENAI_EMBEDDING_MODEL>" }{
"mcpServers": {
"memforge-memory": {
"command": "npx",
"args": [
"-y",
"@memforgeai/memory-service"
],
"env": {
"DATABASE_URL": "<DATABASE_URL>",
"OPENAI_API_KEY": "<OPENAI_API_KEY>",
"OPENAI_BASE_URL": "<OPENAI_BASE_URL>",
"OPENAI_EMBEDDING_MODEL": "<OPENAI_EMBEDDING_MODEL>"
}
}
}
}{
"servers": {
"memforge-memory": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@memforgeai/memory-service"
],
"env": {
"DATABASE_URL": "<DATABASE_URL>",
"OPENAI_API_KEY": "<OPENAI_API_KEY>",
"OPENAI_BASE_URL": "<OPENAI_BASE_URL>",
"OPENAI_EMBEDDING_MODEL": "<OPENAI_EMBEDDING_MODEL>"
}
}
}
}Put your own values in place of
<DATABASE_URL> (PostgreSQL connection string), <OPENAI_BASE_URL> (Embedding API base URL (OpenAI-compatible)), <OPENAI_API_KEY> (Embedding API key), <OPENAI_EMBEDDING_MODEL> (Embedding model name (e.g. BAAI/bge-m3)).
The author’s registry entry marks them as required.
Alternatives to Memforge Memory
same job, measured the same wayPersistent memory for AI coding agents. Local-first SQLite + vector search. 17 MCP tools.
MCP memory server for AI coding assistants. Windows Service, ONNX embeddings, SQLite vector search.
Persistent semantic memory for AI agents using local ChromaDB vector search. No cloud required.
File-first, local-first MCP memory for AI coding assistants: Markdown + YAML, no vector DB.
Memory for AI assistants in one SQLite file: full-text search, no vectors, no embeddings.
Private on-device memory & retrieval for AI assistants — offline vector + lexical search.
Persistent memory and knowledge graph for AI assistants — keyword + vector + graph search.
Persistent knowledge memory for AI agents. Hybrid search, code graph, pgvector.
Memforge Memory — questions
Answers built from our own checks of this server.