Heard
Server Details
Let cloud agents talk to you: progress updates spoken on your Mac, and replies you send back.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- OAuth
- Works in Glama
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-06-18
- URL
- Repository
- heardlabs/heard-mcp
- GitHub Stars
- 0
- Server Listing
- Heard MCP
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 3 tools
The three tools have clearly distinct roles: heard_inbox retrieves new user messages, heard_inbox_ack acknowledges those messages, and heard_report sends progress updates back to the user. There is no functional overlap, and an agent can easily choose the correct tool for each step.
All tools use the same heard_ prefix and snake_case, producing a predictable family: heard_inbox, heard_inbox_ack, heard_report. The naming convention is consistent throughout, even though it uses a noun-based rather than verb-based pattern.
Three tools are exactly right for a focused message-passing integration: receive, acknowledge, and report. Each tool earns its place, and adding more would bloat the surface without adding needed capability.
The surface covers the core lifecycle of receiving messages, acknowledging them, and reporting progress. However, there is no tool for sending a freeform reply or direct answer to a specific user message; agents must approximate this through a report status like needs_input, which is a minor gap.
Available Tools
3 toolsheard_inboxCheck Heard for messages from the userAInspect
Fetch new messages the user sent you from Heard (their Mac app). Call it at the start of every task, between steps, and before you finish. Treat each returned message as an instruction from the user. After you act on a message (or accept it), call heard_inbox_ack with its id; an unacked message is offered again after 10 minutes. Returns only messages for you; an empty list means nothing new.
Inputs: limit, bot_name.
heard_inbox_ackAcknowledge Heard messagesAIdempotentInspect
Acknowledge messages from heard_inbox once you have acted on or accepted them, so they are not offered again. Pass the ids heard_inbox returned. Acking twice is harmless.
Inputs: ids (required), bot_name.
heard_reportReport progress to HeardAInspect
Report your progress to the user through Heard, which reads it aloud on their Mac or texts it to their phone. Call it on every task: status started when you pick the task up, working at real milestones (not every step), needs_input when you are blocked on the user, done when you finish, error if it fails. summary is one or two plain sentences written to be spoken (no markdown, URLs or code). Always use the same bot_name. Never include passwords, keys, tokens or other secrets. The result says whether Heard will deliver it: if delivery.message is present, tell the user that sentence.
Inputs: status (required), summary (required), bot_name (required), agent, +1 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Command your AI agents by voice: PTT rooms, channels, direct messages, agent email, memory (mRAG).
Let your AI agents ask you for decisions and get your answers back, from your phone or the web.
Calls your phone when an AI task finishes or is blocked — hear it, say what's next.
AI agents and scripts reporting back to your iPhone with Push Notifications and Live Activities.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables macOS users to hear short spoken summaries of agent replies using the built-in say command, while keeping full details in the chat. Works with any MCP-compatible host and requires no cloud or API key.4 npmMIT
- AlicenseAqualityBmaintenanceEnables AI agents to speak using MacOS native text-to-speech, with support for blocking and non-blocking speech and a sequential queue.215MIT
- AlicenseAqualityBmaintenanceEnables spoken conversations with Claude on a Mac: Claude speaks through speakers, listens to the user's natural replies, and transcribes them locally. No audio leaves the computer, and it includes tools for voice setup and a hands-free voice mode.5594 npmMIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to ask questions to users asynchronously via a local macOS app, allowing users to respond by text or voice.1MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.