Koboshi

app.koboshiv1.0.0Updated Oct 3, 2026

Save vocabulary from any AI chat; see each word in every language you know; review it with FSRS.

VerifiedStreamable HTTPWeb executableKnowledge & Memory

Overview

AI-generated overview

Lets an assistant save vocabulary words from a chat and review them later with spaced repetition across the languages you know.

What it does
Koboshi is a remote MCP endpoint for collecting vocabulary while you chat with an AI assistant. According to its description, it saves words from any AI chat, shows each word in every language you know, and supports review using FSRS, a spaced-repetition scheduling algorithm. The manifest lists no tools, so the exact operations available are not documented here.
When to use it
Worth considering if you study languages and want words you encounter during AI conversations captured automatically and reviewed on a spaced-repetition schedule. Less useful if you do not keep vocabulary lists or do not want chat content saved to a third-party service.
Requirements
A remote MCP endpoint at using streamable HTTP. The manifest declares no authentication, environment variables, or headers, and no local package or runtime is needed. A Koboshi account may be required in practice, but that is not stated in the provided information.
Before you install
The server is remote, so vocabulary saved from your chats is sent to and stored by a third party. The manifest declares no authentication, which means access control is unclear; check how the service identifies you and who can read your saved words before connecting it to conversations containing sensitive content.

Installation

In SourceWeft

  1. Open Koboshi in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "koboshi": {
      "type": "http",
      "url": "https://koboshi.app/mcp"
    }
  }
}

Tools

0
Tool metadata has not been indexed yet.

Version history

1
  1. v1.0.0LatestOct 3, 2026