total-recall-gemini
Weaving long-term memory and semantic intelligence into the Gemini CLI. A Synaptic Weave Utility.
Open source Open in the app JSON README (API)
About
Weaving long-term memory and semantic intelligence into the Gemini CLI. A Synaptic Weave Utility.
Details
- Kind
- Plugins
- Topic
- AI, RAG & memory
- Publisher
- synaptic-weave
- Origin
- gemini
- Category
- ferramentas
- Version
- 0.2.0
- Forks
- 1
- Last push
- 2026-05-14T20:19:53Z
- Repository state
- ativo
- Language
- Python
- License
- MIT
- Added
- 2026-08-30 14:13:39
- Updated
- 2026-08-30 14:13:39
- Origin id
synaptic-weave/total-recall-gemini
README
# Total Recall (Gemini CLI Edition) > **Weaving long-term memory and semantic intelligence into the Gemini CLI.**   "Total Recall" is a privacy-first, zero-knowledge memory system for your Gemini CLI conversation history. It indexes your past interactions using high-performance semantic search, allowing you to build a persistent "Tapestry" of knowledge across all your AI tools. --- ## 🚀 Features - **Conversational Recall**: Just ask Gemini "Where were we?" or "Do you remember the auth bug?" and it will autonomously search your history using its **Agent Skill**. - **Hybrid Architecture**: Combines **MCP Tools** for reliable data retrieval with **Agent Skills** for seamless, natural language interaction. - **Project Status (`/status-report`)**: Instantly retrieve and summarize the last 10 messages from your current project context. - **Performance Optimized**: - **Surgical Ingestion**: Only indexes the current session's changes on every turn. - **Background Processing**: Hooks run in the background to ensure zero lag in the CLI. - **Lazy Loading**: AI models are only loaded when new data needs to be indexed. - **Persistent Sync**: Automatically catches up on all missed history, even after long gaps between sessions. - **Zero-Knowledge Architecture**: Every message chunk and project tag is encrypted locally using **AES-256-GCM** before being stored. - **High Performance**: Built on **APSW** and **`sqlite-vec`**, providing sub-millisecond vector retrieval for massive databases. - **Shared Enclave**: Stores your Master Key in the OS-native keychain for a seamless, password-free daily experience. - **Cross-Tool Synergy**: 100% schema-compatible with the [Claude Code version](https://github.com/aguywithcode/total-recall). ## 📦 Installation ### 1. Install through Gemini CLI ```bash gemini extensions install https://github.com/synaptic-weave/total-recall-gemini ``` ### 2. Automatic Dependency Setup Total Recall includes a "Lazy Bootstrapper." On its first run, the extension will automatically check for and install all required Python libraries (including `apsw`, `sqlite-vec`, and `mcp`) via pip. No manual installation is required. --- ### Alternative: Manual Installation (for development) ```bash git clone https://github.com/synaptic-weave/total-recall-gemini.git cd total-recall-gemini pip install -r requirements.txt gemini extensions link . ``` ## 🛠 Usage ### 🗣️ Conversational Interaction (Recommended) You can speak naturally to Gemini. It will use its memory tools when appropriate: - *"Where were we on this project?"* - *"What was the last thing we did here?"* - *"Can you recall our work on the database migration?"* ### ⚡ Explicit Shortcuts - **Search History**: `/recall <query>` - **Project Summary**: `/status-report` ## 🔐 Security & Privacy ### Wrapped Key Model - **Master Key (MK)**: A random 256-bit key that encrypts all your data. - **Master Password**: Used to "wrap" the MK. You are prompted for this once to unlock the session. - **Recovery Phrase (BIP39)**: A 12-word phrase generated during setup. This is the **only** way to recover your data. - **Keychain Storage**: Once unlocked, the MK is stored in your system's secure enclave (macOS Keychain / Windows Credential Manager) for a seamless experience. ## 📄 License MIT --- *A Synaptic Weave Utility*