Ai Context Flow

Name: Ai Context Flow
Author: Web3-Plurality

@Web3-Plurality

Repository

About Ai Context Flow

Universal memory for AI agents and tools. Save, organize and search context on any AI tool or platform.

#ai-memory #ai-context #universal-memory

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "ai-context-flow": {
      "type": "streamable-http",
      "url": "https://app.plurality.network/mcp"
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Ai Context Flow?

Ai Context Flow is an OAuth-secured Model Context Protocol (MCP) server that gives any MCP-compatible AI client (Claude Code, Claude Desktop, Cursor, ChatGPT, etc.) read and write access to a user's Plurality memory — documents, notes, and files stored across memory buckets.

How to use Ai Context Flow?

Install dependencies with uv sync, configure the .env file, start Hydra+Traefik via Docker Compose, then run uv run uvicorn main:mcp_server --host 0.0.0.0 --port 5051. In production, point your MCP client to https://app.plurality.network/mcp and authenticate via OAuth 2.1 + PKCE (interactive) or a Personal Access Token (headless). For Claude Desktop, use the mcp-remote bridge; for Claude Code, run claude mcp add --transport http plurality-memory [URL].

Key features of Ai Context Flow

OAuth 2.1 + PKCE or Personal Access Token authentication
7 MCP tools: bucket listing, item search, content read/write, conversation save
Semantic memory search across buckets with relevance scoring
JWT validation via Hydra JWKS keys (cached 1 hour)
Streamable HTTP transport with Dynamic Client Registration
Local development support with Traefik reverse proxy

Use cases of Ai Context Flow

Save conversation history to a memory bucket for later recall
Semantic search across stored documents and notes
Read full content of stored items with pagination
Create new memory buckets to organize saved content
Headless agents (CI runners, n8n, LangChain) using PAT authentication

FAQ from Ai Context Flow

Which MCP clients are supported?

Claude Desktop (via OAuth or mcp-remote bridge), Claude Code (terminal and VSCode extension), ChatGPT (paid plan, OAuth), Cursor, and any client supporting streamable HTTP and DCR.

What authentication methods are available?

Interactive clients use OAuth 2.1 + PKCE (browser login). Headless agents and custom integrations use a Personal Access Token (PAT), which requires a paid plan and is managed from the dashboard.

Does Ai Context Flow require a paid plan?

OAuth-based access is available on free plans for development via the Desktop app with mcp-remote. PATs and the ChatGPT connector require a paid plan (Pro, Max, Team, Enterprise).

What are the runtime requirements?

Python 3.11+, the uv package manager, Docker and Docker Compose (for Hydra and Traefik), and running Plurality backend services (API Gateway and Vector Service).

How are tokens validated?

The MCP server validates JWT access tokens (RS256, 15‑minute TTL) locally using Hydra’s JWKS public keys, verifying the mcp:tools scope. Tokens are cached for 1 hour.

Comments

More Other MCP servers

Unity MCP ✨

justinpbarnett

Unity MCP acts as a bridge between AI assistants and your Unity Editor. Give your LLM tools to manage assets, control scenes, edit scripts, and automate tasks within Unity.

12.2K