langchain-perseus-vault 0.1.0

Perseus Vault persistent, local, encrypted memory for LangChain — tools and a retriever backed by the Perseus Vault MCP engine.

langchain-perseus-vault

Persistent, local-first, encrypted memory for LangChain, backed by Perseus Vault (formerly "Mimir"/"Mneme") — an open-source (MIT) memory engine with FTS5 + dense hybrid search and optional AES-256-GCM encryption, exposed over the Model Context Protocol (MCP) stdio transport.

It gives a LangChain agent durable memory that survives across runs and processes, stored in a single local SQLite file you control — no external service, no cloud.

What you get

This package wraps Perseus Vault using the modern langchain-core interfaces:

• create_perseus_vault_tools(client) — a pair of StructuredTools (perseus_vault_remember / perseus_vault_recall) you give to an agent so it can manage its own long-term memory via tool calls. This is the current-recommended LangChain pattern (the legacy Memory / ConversationBufferMemory classes are deprecated).
• PerseusVaultRetriever — a BaseRetriever returning Documents, for drop-in use in RAG chains and anywhere LangChain accepts a retriever (.invoke(query)).
• PerseusVaultClient — the low-level MCP stdio client, if you want direct access.

Prerequisite: the perseus-vault binary

This package talks to a local perseus-vault executable via JSON-RPC over stdio. You must have it installed:

• Download a release from https://github.com/Perseus-Computing-LLC/perseus-vault/releases, or build from source (cargo build --release), and put perseus-vault on your $PATH.
• Or pass an absolute path: PerseusVaultClient(perseus_vault_binary="/path/to/perseus-vault").

On Windows the binary may be named perseus-vault.exe; ensure its directory is on PATH, or pass the full path. (Some installs also ship a mimir compat symlink, but perseus-vault is the canonical name.)

Install

pip install langchain-perseus-vault

Usage

As agent tools

from langchain_perseus_vault import PerseusVaultClient, create_perseus_vault_tools

client = PerseusVaultClient(db_path="~/.langchain/mimir.db")
tools = create_perseus_vault_tools(client)  # [perseus_vault_remember, perseus_vault_recall]

# Bind to any tool-calling model / agent:
from langchain.chat_models import init_chat_model

llm = init_chat_model("anthropic:claude-sonnet-4-5")
llm_with_memory = llm.bind_tools(tools)

resp = llm_with_memory.invoke("Remember that my favorite language is Rust.")
# ... the model will call perseus_vault_remember; execute the tool call as usual.

As a retriever

from langchain_perseus_vault import PerseusVaultClient, PerseusVaultRetriever

client = PerseusVaultClient(db_path="~/.langchain/mimir.db")
client.remember("The capital of France is Paris.")

retriever = PerseusVaultRetriever(client=client, k=5)
docs = retriever.invoke("What is the capital of France?")
print(docs[0].page_content)  # -> "The capital of France is Paris."

Direct client

from langchain_perseus_vault import PerseusVaultClient

client = PerseusVaultClient(db_path="~/.langchain/mimir.db")
client.remember("Project deadline is July 15.", tags=["project", "deadline"])
items = client.recall("when is the deadline")
print(items[0]["text"])

How it works

PerseusVaultClient spawns perseus-vault --db <path> as a subprocess and speaks JSON-RPC 2.0 (MCP) over its stdin/stdout. A background reader thread and a lock make calls thread-safe and timeout-bounded. Memories are stored via perseus_vault_remember and retrieved via perseus_vault_recall.

License

MIT © 2026 Perseus Computing LLC