Mneme
中文
Documentation
Getting startedConfigurationTUI commandsSupported files简体中文View on GitHub ↗

Mneme/Guide · 8 min read

Guide · 8 min read

Getting started

Install Mneme, create a local index, and launch the bilingual terminal UI. The smallest useful path starts with files you already own.

Before you begin

Mneme requires Python 3.10 or newer and an OpenAI-compatible endpoint when you use answer generation, query decomposition, or Graph RAG entity extraction. Indexing and retrieval can be prepared locally.

A useful boundary

Do not put API keys, passwords, or other secrets inside the directory you index. Retrieved snippets are sent to the endpoint you configure only when an LLM-backed feature is used.

Install

git clone https://github.com/realhenrylan/mneme.git\ncd mneme\npython -m venv .venv\n\n# Windows PowerShell\n.venv\Scripts\Activate.ps1\n\n# macOS / Linux\nsource .venv/bin/activate\n\npython -m pip install -e ".[dev]"

Configure an endpoint

Copy the example environment file and set the minimum connection values. The onboarding wizard can also collect these settings at first launch.

copy .env.example .env       # Windows PowerShell\n# cp .env.example .env       # macOS / Linux\n\nAPI_KEY=sk-your-api-key-here\nBASE_URL=https://api.deepseek.com/v1\nLLM_MODEL=deepseek-chat

Non-local endpoints must use HTTPS by default. Local loopback endpoints may use HTTP; an insecure non-local endpoint requires an explicit override.

Launch the TUI

python -m tui

The TUI supports streaming chat, Standard RAG and Graph RAG modes, file management, directory watching, settings, and source display.

Where to go next

GUIDEConfigurationFEATUREHybrid RetrievalGUIDETUI commandsFEATURESafety boundaries

Search Mneme documentation

Getting startedInstall and launch Mneme.Hybrid RetrievalFind meaning and exact terms together.Graph RAGFollow relationships across documents.SafetyUnderstand what stays local and what can leave.