OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Memory & RAG

Yantrikdb — yantrikos

YantrikDB is an embeddable Rust memory engine for AI agents, with Python bindings. It records and recalls memories, manages temporal decay and consolidation, detects contradictions, and supports knowledge-graph relationships and HNSW approximate-nearest-neighbor retrieval.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Yantrikdb — yantrikos — OpenViST project profile in Agent Memory & RAG.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use YantrikDB when an agent needs a memory-management layer that combines recall with decay, consolidation, contradiction detection or knowledge-graph relationships, rather than a store that only retains vector records.

Who it is for

  • Developers adding persistent memory to AI agents.
  • Python or Rust developers who want to record and recall agent knowledge within an application.
  • Teams configuring an agent through an MCP client.

Documented uses

  • Record facts and decisions so an agent can retrieve them across sessions.
  • Recall semantically relevant memories using a query.
  • Organize related entities as graph relationships.
  • Consolidate stored memories and detect conflicting information.
Before you choose

Requirements & limits

The repository identifies YantrikDB as Apache-2.0 licensed. The README documents installation through pip for the Python library and MCP package.

  • The core library is embeddable and written in Rust, with Python bindings.
  • A new default Python store fetches and caches a 256-dimensional embedding model on first use; an offline fallback uses a bundled 64-dimensional model.
  • A store keeps its existing embedding dimension when reopened.
  • The README warns that using a second SQLite library on the same store in the same process can corrupt it; it recommends the engine API or a separate process.
  • HTTP gateway access is described as a separate yantrikdb-server project, not as the core library.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up a Python memory store and try recall

Create a file-backed YantrikDB store, add sample memories, retrieve a relevant one, and exercise graph and consolidation calls.

Before you start

  • Python environment with the yantrikdb package installed using pip.
  • Network access on first use for the default model, unless relying on the bundled offline fallback.

Steps

  1. Install the package with the command below.

    Documented command
    pip install yantrikdb

  2. Open a store with yantrikdb.YantrikDB.with_default("memory.db").

  3. Add text using record(), then call recall() with a query and a top_k value.

  4. Use relate() and get_edges() for graph relationships, then call think() to consolidate and detect conflicts.

  5. Close the store with close().

Expected outcome

A file-backed store containing recorded memories, with documented calls for recall, graph relationships and memory consolidation.

Guide limitations

  • The default model for a new store is fetched once and cached; if it cannot be fetched, the bundled 64-dimensional model is used with a warning.
  • An existing database reopens at its existing embedding dimension.
  • Do not open the same store through a second SQLite library in the same process; use the engine API or a separate process.
Documentation 1 ↗
Source-based practical guide

Configure YantrikDB through an MCP client

Add the documented YantrikDB MCP command to an MCP client configuration.

Before you start

  • Python package installation support.
  • An MCP client that accepts a server command configuration.

Steps

  1. Install the MCP package with the command below.

    Documented command
    pip install yantrikdb-mcp

  2. Add a yantrikdb entry to the MCP client configuration with command set to yantrikdb-mcp.

Expected outcome

The client configuration contains the YantrikDB MCP server command for agent memory integration.

Guide limitations

  • The README points to the separate yantrikdb-mcp repository for full documentation.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
yantrikos/yantrikdb · Official homepage title · retained in its source language
How we review sourcesSuggest a correction

Related options

yantrikdb-server — A separate server project that wraps the engine with an HTTP gateway and MCP server interface.

yantrikdb-mcp — A separate package whose command can be configured as an MCP server for agent integrations.