Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Bell call notify_done with summary 'Task complete'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Desktop popup notifications that ring when Claude pauses on you.
Claude Code: fires a popup on Stop (turn finished, only if the turn took >30s), Notification (Claude is idle or asking for permission), and SubagentStop.
Claude Desktop: exposes an MCP tool notify_done(summary) that Claude calls when it finishes a task.
Clicking the popup dismisses it and can bring the relevant Claude, VS Code, or terminal window to the front.
Bell records local notification history in ~/.bell/history.sqlite3 using SQLite. Each row includes timestamp, source, title, message, and focus target.
Popups center on the monitor containing the current foreground window. If monitor detection fails, Bell falls back to the primary-display behavior.
cd "C:\Users\<you>\Desktop\projects\bell"
py -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -e .Related MCP server: Notifications MCP Server
python -m bell.notify "Bell" "Hello from Bell" none
echo '{"hook_event_name":"Stop","session_id":"test-session","cwd":"C:\\tmp","transcript_path":"","stop_hook_active":false}' | python -m bell.hook
python -m bell.codex_notify turn-ended
python -m bell.doctorDo not wire this automatically from the repo build. In a later approved step, merge hook commands into the relevant Claude settings file.
The MCP entry point is:
python -m bell.mcp_serverCodex CLI/App can call Bell through its notify config:
notify = ["C:\\Users\\<you>\\Desktop\\projects\\bell\\.venv\\Scripts\\pythonw.exe", "-m", "bell.codex_notify"]Bell only opens a popup for Codex completion notifications: either the turn-ended wrapper event or a JSON payload with type = "agent-turn-complete". Other Codex notify invocations are logged to ~/.bell/bell.log and skipped without writing history.
Show a Bell popup when a task finishes; keep summary short, about 120 characters.
| Name | Required | Description | Default |
|---|---|---|---|
| session | No | ||
| summary | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Describes the output as a Bell popup, but no annotations exist to pre-declare safety or side effects. The description adds the character constraint but lacks details like dismissal behavior, persistence, or whether it blocks. For a simple notification, this is moderately transparent.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single sentence, front-loaded with action, no waste. Essential information is conveyed efficiently.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Tool is simple, and output schema likely covers return values. However, the description omits session parameter context and does not clarify if any side effects (e.g., persistent notifications) exist. Adequate but not fully complete given no annotations.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 0%, so description must compensate. It explains 'summary' with a character limit but does not define its role (e.g., message text). 'Session' parameter is completely unaddressed, leaving ambiguity. This partially adds meaning but insufficiently for two parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Clearly states the tool shows a Bell popup when a task finishes, with a specific verb ('Show') and resource ('Bell popup'). Distinguishes from any potential sibling by its unique action, though no siblings are listed.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Provides a guideline to keep the summary short (~120 characters), which implies appropriate usage context for concise notifications. No explicit when/when-not guidance, but the tool's purpose is self-evident given no siblings.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
With only one tool, there is no possibility of confusion. The tool's purpose is clearly defined and distinct.
A single tool name cannot be inconsistent. 'notify_done' follows a clear verb_noun pattern.
One tool for a simple notification purpose is reasonable, though the server could potentially benefit from additional tools like notify_error or notify_progress.
The single tool covers the core use case of notifying task completion, but lacks other common notification types, making it somewhat incomplete for broader notification needs.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Let your AI agent notify you by email, Slack, Discord, or webhook. One tool: send_notification.
Push notifications for AI agents - send instant iPhone notifications from any MCP client.
Create Tasks and run Workflows in Bleep from Claude, ChatGPT, and other AI assistants.
Reach your own phone from an AI agent: notifications, approval questions, reminders, ring, files.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/YongyangFu/bell'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
