Delive, developed by Ximilalaxiang, is a free MCP server tool built with TypeScript and Electron, running on Node.js. It functions as a desktop transcription workspace, capturing system audio for real-time processing or handling uploaded audio/video files offline. Delive routes audio through various local or cloud-based Automatic Speech Recognition backends. It provides interfaces for AI agents and clients, enabling advanced analysis of transcription sessions. Distributed as binary packages for Windows, macOS, and Linux, it also supports auto-updates.

What does Delive connect to?

Delive connects to a computer's system audio to capture playback for transcription. It routes this audio through multiple Automatic Speech Recognition (ASR) backends, which can be either local or cloud-based. For offline transcription, Delive connects to cloud engines to process uploaded audio and video files. It also exposes an MCP server, a local REST API, and a real-time WebSocket, designed for compatibility with AI agents and AI clients that adhere to the Model Context Protocol standard.

Core features and tools

Delive functions as a desktop transcription workspace for system audio. It supports multi-provider ASR backends and offline transcription of uploaded media. It exposes an MCP server for AI agents, a local REST API, and a real-time WebSocket. A specific callable tool is delive-transcript-analyzer, which offers functions for searching, generating summaries, extracting insights, and correcting Delive transcription sessions. The platform is designed to work with AI agents and clients using MCP SDKs.

Who is it for?

Delive is for users needing a desktop transcription workspace for system audio, such as for meetings, lectures, or general computer audio. It also targets users who want to use AI for reviewing and processing transcripts. This includes individuals who require real-time transcription with an AI Review Desk for summary, chat, Q&A, and mind mapping. It also suits those needing offline transcription of audio/video files. Skip this if you do not require system audio transcription or AI-powered transcript analysis.

How to get started

Access is granted through optional Bearer token authentication, which is disabled by default. The MCP server supports transport via a local REST API and real-time WebSocket.

Use cases

• Real-time transcription of system audio with an AI Review Desk for summary, chat, Q&A, and mind mapping.
• Offline transcription of audio and video files using various cloud engines.
• Programmatic integration with AI agents to analyze transcription sessions, perform searches, generate summaries, extract insights, and correct transcripts.

Delive offers a desktop environment for transcribing system audio and uploaded media, integrating with AI for advanced analysis. Its MCP server, REST API, and WebSocket interfaces enable connections with AI agents and clients. The platform supports real-time transcription and offline processing, providing tools for summary, search, and correction of transcripts. Its optional Bearer token authentication allows for secure, configurable access.