Webear
Let your agent hear and reason about live browser or app audio while you debug or build multimodal features.
Overview
WebEar is an MCP server for the Build phase that captures live audio from your dev server and returns agent-ready analysis and descriptions via codedswitch.
What is this MCP server?
- stdio npm MCP package webear v1.2.4
- Live audio capture via WebEar middleware on your dev server (WEBEAR_BASE_URL)
- Analysis and natural-language descriptions powered by codedswitch.com (CODEDSWITCH_API_KEY required)
- Designed to give Claude Code and similar assistants contextual hearing during coding sessions
- Registry version 1.2.4
- Transport: stdio via npm identifier webear
- Required secret env: CODEDSWITCH_API_KEY
What problem does it solve?
Your coding agent cannot hear what users hear in the browser or app, so debugging audio UX and describing live sound requires constant manual copy-paste and guesswork.
Who is it for?
Indie developers building web audio, notifications, or accessibility features who want Claude Code to reason about live sound during local dev.
Skip if: Teams that only need offline file transcription or production monitoring without a dev-server middleware layer.
What do I get? / Deliverables
After install, your assistant can request captures and structured audio descriptions tied to your running dev environment during implementation and debugging.
- stdio MCP tools for live audio capture and description
- Agent-readable summaries of audio heard through your dev stack
Recommended MCP Servers
Journey fit
WebEar extends the coding agent with real-time audio capture and analysis during product construction, not during launch or growth analytics. It wires middleware on your dev server into MCP so assistants get structured audio descriptions—classic agent-tooling for local development workflows.
How it compares
Dev-server audio MCP bridge, not a standalone speech API or monitoring SaaS.
Common Questions / FAQ
Who is WebEar for?
Solo and small-team builders using AI coding agents who need live audio context from a local or staging app while implementing or debugging sound-related features.
When should I use WebEar?
Use it during Build when WebEar middleware is mounted on your dev server and you want MCP tools to capture and describe audio instead of describing problems manually.
How do I add WebEar to my agent?
Install the npm package webear, configure WEBEAR_BASE_URL to your middleware URL and CODEDSWITCH_API_KEY, then register the stdio MCP server in Claude Code, Cursor, or another MCP-capable client.