# AI Context Flow

Gives MCP clients persistent memory: documents, notes and chats sit in topic buckets and come back through semantic search. Useful when several agents need the same context.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 2
- Author: Web3-Plurality
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/Web3-Plurality/plurality-mcp-server
- Updated: 2026-08-30
- HTML: https://affpapa.org/en/nevibecode/mcp/ai-context-flow

## What it does

The server connects over Streamable HTTP and authenticates either through OAuth 2.1 with PKCE or with a Personal Access Token. Seven tools cover the whole loop: list buckets, list what a bucket holds, search across buckets, read one item in full, save text, save a conversation, create a bucket. Matching happens on meaning rather than exact wording, so you can phrase a query however you like.

Buckets can be shared with other people, which turns your saved context into something a team works from. Headless agents and CI runners get a token from the dashboard and send it as `Authorization: Bearer plur_pat_...`. Tokens need a paid plan and are stored hashed.

Run an agent with no browser and OAuth has nowhere to go — the one-time sign-in is for interactive clients, while CI and custom integrations are stuck with a PAT. Also worth knowing: the endpoint is `https://app.plurality.network/mcp`, nothing runs on your machine, so that host has to be reachable.

## Requirements
- Remote server: `https://app.plurality.network/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http ai-context-flow https://app.plurality.network/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add ai-context-flow --url https://app.plurality.network/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "ai-context-flow": {
            "url": "https://app.plurality.network/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "ai-context-flow": {
            "type": "http",
            "url": "https://app.plurality.network/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http ai-context-flow https://app.plurality.network/mcp
```

## When to use it
- pull up a note by meaning instead of an exact keyword
- file a chat transcript into its own bucket
- hand the same context to more than one agent
- sort documents and files into topic folders
- let someone else read part of your saved context

## Example prompts
- List all my memory buckets
- Search my memory for CI setup notes
- Save this conversation to the `backend` bucket

## Similar
- [openviking-memory](https://affpapa.org/en/nevibecode/skills/openviking-memory.md) — skill · 504 installs
- [saving-workspace-context](https://affpapa.org/en/nevibecode/skills/saving-workspace-context.md) — skill · 495 installs
- [shodh-memory](https://affpapa.org/en/nevibecode/mcp/shodh-memory.md) — MCP server · ★ 316
- [aidex](https://affpapa.org/en/nevibecode/mcp/aidex.md) — MCP server · ★ 46
- [context-mem](https://affpapa.org/en/nevibecode/mcp/context-mem.md) — MCP server · ★ 19
- [code-context](https://affpapa.org/en/nevibecode/mcp/code-context.md) — MCP server · ★ 32

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md