# FeedKeeper MCP server

> Lets an AI client read, search, bookmark, and organize RSS and Atom feeds through a remote MCP endpoint.

- Repository: https://github.com/visualfusion/feedkeeper
- Page: https://hostedmcp.org/servers/visualfusion/feedkeeper
- Transport: stdio, remote
- Stars: 1
- License: MIT
- Latest version: 1.0.3

## Install

### Claude Code

```sh
claude mcp add feedkeeper -- npx -y feedkeeper
```

### Cursor (`~/.cursor/mcp.json`)

```json
{
  "mcpServers": {
    "feedkeeper": {
      "command": "npx",
      "args": [
        "-y",
        "feedkeeper"
      ]
    }
  }
}
```

### VS Code (`.vscode/mcp.json`)

```json
{
  "servers": {
    "feedkeeper": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "feedkeeper"
      ]
    }
  }
}
```

### VS Code CLI

```sh
code --add-mcp '{"name":"feedkeeper","type":"stdio","command":"npx","args":["-y","feedkeeper"]}'
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "feedkeeper": {
      "command": "npx",
      "args": [
        "-y",
        "feedkeeper"
      ]
    }
  }
}
```

### Codex

```sh
codex mcp add feedkeeper -- npx -y feedkeeper
```

### Codex config.toml (`~/.codex/config.toml`)

```toml
[mcp_servers.feedkeeper]
command = "npx"
args = ["-y", "feedkeeper"]
```

### Gemini CLI

```sh
gemini mcp add -s user feedkeeper npx -- -y feedkeeper
```

### Gemini settings.json (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "feedkeeper": {
      "command": "npx",
      "args": [
        "-y",
        "feedkeeper"
      ]
    }
  }
}
```

## Tools (34)

- `list_feeds`
- `get_feed`
- `subscribe_feed`
- `discover_feeds`
- `save_to_inbox`
- `unsubscribe_feed`
- `update_feed`
- `list_folders`
- `create_folder`
- `rename_folder`
- `delete_folder`
- `move_feed_to_folder`
- `refresh_feed`
- `get_digest`
- `get_overview`
- `get_new_items`
- `list_items`
- `get_item`
- `fetch_full_text`
- `search_items`
- `mark_read`
- `mark_unread`
- `mark_all_read`
- `bookmark_item`
- `unbookmark_item`
- `list_muted_keywords`
- `add_muted_keyword`
- `unmute_keyword`
- `export_opml`
- `import_opml`
- `get_edition`
- `get_edition_candidates`
- `publish_edition`
- `dismiss_edition`

## FAQ

### How can I use FeedKeeper with AI agents like Claude or ChatGPT?

You can connect MCP clients directly to the `/mcp` endpoint using OAuth sign-in or personal access tokens to read, search, or curate feeds.

### Does FeedKeeper require an external database like PostgreSQL?

No, it is designed to be lightweight and runs using a single Node process and a local SQLite database.

### Can I use FeedKeeper on a Raspberry Pi?

Yes, it is optimized for low-power hardware and can be run on a Raspberry Pi using Docker or Node.js.