Engram

Engram is an advanced AI Memory Engine designed to facilitate knowledge management through a comprehensive approach that combines graph storage, semantic search, logical reasoning, and continuous learning, all encapsulated within a single binary. This innovative system utilizes a single .brain file for the entire knowledge base, simplifying management and ensuring seamless portability.

Key Features:

• Single Binary: Engram operates without runtime dependencies, making installation straightforward with no requirement for Docker or cloud infrastructure.
• Compact Knowledge Base: The complete knowledge base is contained within a single .brain file, allowing for quick backup and migration simply by copying or moving the file.
• All-in-One Functionality: Complete functionality is built in with no need for external databases.
• Hybrid Search Capabilities: Engram offers robust hybrid search options, integrating BM25 for full-text search, HNSW for vector similarity, and bitmap filtering for enhanced performance.
• Confidence Lifecycle Management: Knowledge within Engram is managed through a confidence lifecycle that strengthens knowledge with confirmations, diminishes it over time, and corrects inconsistencies when contradictions arise.
• Inference Engine: Supports forward and backward chaining, rule evaluations, and transitive reasoning for logical deductions.
• Knowledge Mesh: Offers peer-to-peer synchronization between Engram instances, enhanced by an ed25519 identity verification and trust scoring system.
• User-Friendly Interface: An integrated web UI provides a visual graph explorer, search functionalities, natural language query capabilities, and a learning dashboard.
• Multiple APIs: Access through various APIs including HTTP REST, MCP, gRPC, and LLM tool-calling for versatile integration options.

Usage Examples:

To create a new knowledge base, simply run:

engram create my.brain

To store knowledge, use:

engram store "PostgreSQL" my.brain
engram store "Redis" my.brain
engram relate "PostgreSQL" "caches_with" "Redis" my.brain

To query the knowledge base, the command is:

engram query "PostgreSQL" 2 my.brain

Engram also supports powerful searches:

engram search "database" my.brain
engram search "confidence>0.8" my.brain
engram search "type:server AND confidence>0.5" my.brain

Engram provides full documentation for both CLI and HTTP API functionalities, making it easy to integrate into existing workflows.

Use Cases:

Engram's versatility allows it to be used in various scenarios including building knowledge graphs from Wikipedia data, document ingestion with entity extraction, IT support knowledge bases, threat intelligence mapping, and OSINT tasks among others.

Engram is developed for high performance in handling a range of knowledge-based tasks, utilizing techniques such as confidence management and hybrid search to ensure efficient data retrieval and insights.