Yantrikdb Server by Yantrikos is a free MCP server and network database built with Rust, designed as a cognitive memory database for AI agents. It provides a persistent and intelligent memory layer for AI-native systems. Distributed as a single Rust binary and available via Docker, it caters to developers building solutions for "chatty agent workloads." The platform integrates with various AI clients, offering a specialized backend for managing agent context and knowledge.

What does Yantrikdb Server connect to?

Yantrikdb Server functions as an MCP server and a multi-tenant HTTP API network database. It is designed to connect directly to AI agents, providing them with a cognitive memory layer. The server supports HTTP and a binary wire protocol for communication. Additionally, it offers Server-Sent Events (SSE) for IDE integrations. The yantrikdb-mcp component, a Python package, can operate over stdio or HTTP/SSE. Confirmed compatible AI clients include Claude Code, Cursor, Windsurf, Copilot, Continue, Claude Desktop, Hermes Agent, and Prime Agent.

Core features and tools

The yantrikdb-mcp server exposes 16 callable tools for AI agents. These include remember, recall, forget, correct, think, graph, conflict, trigger, session, temporal, procedure, category, personality, stats, memory, and skill. The underlying engine provides a more detailed API with operations like record, record_batch, relate, get_edges, scan_conflicts, session_start, session_end, stale, upcoming, archive, hydrate, decay, evict, list_memories, and stats.

Who is it for?

Yantrikdb Server targets AI agents and developers creating AI-native systems, especially those managing "chatty agent workloads." It serves as a foundational component for applications requiring agents to maintain persistent knowledge and context across interactions. The platform is suitable for environments where AI agents need to remember past decisions, detect contradictions, and manage their own memory autonomously. This is not the right fit if your application does not involve AI agents or require a cognitive memory layer.

How to get started

The server authenticates using a Bearer token. Access is granted through per-database access control, requiring a token to be generated for a specific database.

Use cases

• Providing persistent cognitive memory for AI agents to remember decisions, auto-recall context, and detect contradictions across sessions.
• Performing autonomous cognitive memory management by consolidating duplicate memories, detecting contradictions, and fading stale memories over time to improve recall quality.
• Enabling multi-agent shared memory and replication through cluster deployment with YRP native replication, allowing multiple agents or machines to share a consistent memory state with high availability.

Yantrikdb Server offers a specialized solution for AI agent memory management, providing persistence and advanced cognitive functions. Its multi-tenant HTTP API and support for various AI clients make it a adaptable choice for AI-native development. The system addresses the need for agents to maintain context, manage memory, and operate in shared environments.