📡 MCP Radar

bestnite/engram

Self-hosted, multi-user spaced-repetition (SRS) service: FSRS v6 scheduling, 10 card types, Markdown/TeX, REST API and built-in MCP server. Written in Go.

⭐ 5 stars
⑂ 0 forks
Go
momentum ▲ 8.3
created 2026-10-03
on radar since 2026-10-06
anki-alternativeflashcardsfsrsgogolanglearningmcpmemoryoidcpostgresqlpwarest-apiself-hostedspaced-repetitionsqlitesrs
View on GitHub ↗

About engram

Engram is a self-hosted, multi-user, web-first spaced-repetition (SRS) service. Write cards in the browser, or let a script or agent push them in through the API or the built-in MCP server. Scheduling uses FSRS v6.

Live demo: (registration is closed; sign in with the test account test / testdemo)

---------------------- ----------------------

Scheduling uses FSRS v6. Desired retention defaults to 0.90; learning steps, maximum interval and fuzz are all adjustable. Once you have some review history, you can retrain the parameters from it: trigger it on the preset page, it runs as a background job, and you can revert if you don't like the result.

Ten card types are built in: cloze, typed, numeric, single- and multiple-choice, true/false, and more. Card bodies support Markdown and TeX (self-hosted MathJax 3), and HTML goes through an allowlist. Uploaded images are content-addressed.

Everyone gets their own account and their own scheduling state; deck content is shared. Decks have three roles — owner, editor, reader — and you can hand out share links or clone a deck to your own account. A single deck exports to an .edeck package for backup, migration or offline handover.

The REST API lives under /api/v1, the built-in MCP server at POST /mcp (HTTP only, no stdio). The two are equivalent: both authenticate with a per-user API key and call the same service methods, so validation and scheduling rules cannot diverge. Keys carry scopes (read, write, review, keys, admin) and are managed under Settings in the browser; request and response schemas live in schema/.

From the project README.

Maintaining this server?

Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:

[![On MCP Radar](https://mcp.liqiwa.com/badge.svg)](https://mcp.liqiwa.com/s/bestnite--engram.html)

Related MCP servers

sf-shenfeng/learn-shell

An agent-native teaching OS — a classroom that remembers, for you and your AI. MCP server + self-hosted web app.

⭐ 3TypeScript
Reanthesis/use-reanthesis

Connect your AI to Reanthesis — MCP server, agent skills, and plugins so Claude, ChatGPT, and Copilot can build flashcards from your notes and manage your study data. Your AI, your data, your cards.

⭐ 3Shell
Resetnak/cooldeck

Keyboard-first terminal dashboard for Coolify: deployments, live logs across your whole fleet, and an MCP server for your agent

⭐ 9Go
Evermemoai/evermemo

A tiny, universal memory engine for humans and AI agents — CLI, HTTP API, MCP server, and shared memory hub in one small Go binary

⭐ 8Go
YigitCittan/mongorescue

Self-hosted MongoDB backup and disaster recovery: streaming mongodump to disk or S3, safe-clone restores, encryption, alerts, web dashboard and MCP server. Single Go binary.

⭐ 6Go
soma-intelligence/open-health-sync

Self-hosted, open-source API for wearable and health data, in Go. Syncs Apple Health, Health Connect, Oura, WHOOP, Withings, Strava, Fitbit, Dexcom and more into one model, served over REST, FHIR R4,

⭐ 5Go

📬 Get the weekly radar in your inbox

The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.