Persistent memory system for AI coding agents

Engram: Persistent Memory for AI Coding Agents

Engram is an agent-agnostic persistent memory system designed to enhance the capabilities of AI coding agents. Built with Go, SQLite, and FTS5, it provides a robust and efficient solution for storing and retrieving information, enabling AI agents to maintain context and learn over time. This system is crucial for developers building sophisticated AI tools, MCP servers, and agents that require long-term memory and state management.

What Engram Does

Engram acts as a centralized knowledge base for AI coding agents. It allows agents to store, query, and retrieve information in a structured and persistent manner. This means that an AI agent can "remember" past interactions, code snippets, project details, and learned patterns, rather than starting from scratch with each new task. By leveraging SQLite with FTS5 (Full-Text Search), Engram offers powerful and fast search capabilities, making it easy for agents to find relevant information within their memory.

Key Features

• Agent-Agnostic Design: Integrates seamlessly with various AI coding agents without requiring specific modifications to the agent's core logic.
• Persistent Storage: Utilizes SQLite for reliable and long-term storage of agent memories.
• Efficient Search: Implements FTS5 for fast and accurate full-text searching of stored data.
• MCP Server: Includes an MCP server for inter-agent communication and memory access.
• HTTP API: Provides a RESTful HTTP API for programmatic access to memory functions.
• CLI and TUI: Offers both command-line interface (CLI) and terminal user interface (TUI) for direct interaction and management of the memory system.
• Go Implementation: Developed in Go, ensuring performance and concurrency benefits.

Who Engram is For

Engram is an essential tool for AI developers , ML engineers , and anyone building or working with AI coding agents . It is particularly beneficial for projects that involve:

• Developing AI agents that require long-term memory and context.
• Creating complex AI systems that need to learn and adapt over time.
• Building custom MCP servers or agent frameworks.
• Enhancing the productivity and intelligence of AI coding assistants.
• Researchers exploring persistent memory solutions for AI.