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

Magicore

MagiCore is a local-first C# library that adds persistent memory to .NET AI applications and agents, with semantic, temporal and spatial memory features, hybrid retrieval and provider-neutral persistence.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Magicore — 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 MagiCore when a .NET application needs memory beyond the current interaction, especially when scoped semantic search, auditable history, or temporal and spatial records are part of the design.

Who it is for

  • Developers building AI agents or assistants in .NET that need persistent, scoped recall.
  • Teams building simulations or robotics applications that need timestamped spatial observations or action episodes.
  • Developers integrating memory with Microsoft.Extensions.AI, Microsoft.Extensions.VectorData or the .NET MCP SDK.

Documented uses

  • Save user-, session- or agent-scoped information and retrieve relevant context with a query.
  • Combine dense vector retrieval and BM25 keyword scoring, with reranking options.
  • Track memory changes and, with history-capable storage, query past state or roll back selected changes.
  • Represent timestamped 3D observations and recall measured action episodes.
Before you choose

Requirements & limits

The repository identifies the project as Apache-2.0 licensed and documents package installation through NuGet.

  • Basic memory can run in-process using the documented local defaults, without a required hosted service.
  • Durable vector storage is configurable through Microsoft.Extensions.VectorData; the README lists PostgreSQL/pgvector, Qdrant, SQLite, Redis and other providers.
  • Point-in-time reads and rollback require a history-capable store, and provider limitations apply.
  • The documented target frameworks include .NET Standard 2.0 and .NET 8, 9 and 10.

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

Add and search a local memory

Use the documented default service to save a user-scoped fact and retrieve it with a query.

Before you start

  • A .NET project with the MagiCore package installed.
  • C# code using the MagiCore namespace.

Steps

  1. Install the package with dotnet add package MagiCore.

  2. Create a MemoryService instance.

  3. Call AddAsync with a memory and a user ID.

  4. Call SearchAsync with a query and a MemoryFilter for that user.

Expected outcome

The service returns matching memory results; the README example prints the retrieved memory text.

Guide limitations

  • The default store, extractor and lexical embedding generator are local and deterministic; this recipe does not configure durable vector storage or an external model.
Documentation 1 ↗
Evidence and transparency

Sources & review

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