"SearXNG, a privacy-respecting metasearch engine" matching MCP servers:

Enables AI agents to run SearXNG-powered metasearch across multiple engines and categories, with optional vector-based semantic re-ranking of results and Tavily-compatible responses. It also converts many document and media formats—including PDF, Office files, Outlook messages, and audio or YouTube recordings—into text.

Self-host a realtime voice-call companion for coding agents. Exposes an MCP endpoint that agents can poll as an alternate input stream.

Enables voice assistants to find nearby service providers, connect users to live video representatives for natural conversation, and handle appointment booking. It also returns confirmed bookings so the assistant can add them to its calendar.

MCP server for offline speech-to-text and speaker diarization, enabling AI agents to transcribe audio locally without cloud APIs.

Enables MCP-capable chat clients to synthesize speech from text, convert it with a local RVC voice model, and render the result in an inline audio player.

Provides on-device Chinese speech recognition with traditional Chinese (Taiwan) output via MCP server, HTTP API, and CLI, ensuring privacy by processing audio locally without uploading to cloud services.

Enables local audio and video transcription via the audiototext engine, returning detected language, full text, and timestamped segments.

Enables AI agents to synthesize natural speech using either platform system voices or premium OpenAI TTS, with automatic engine selection and graceful fallback.

Enables compatible local MCP hosts to inspect and control a local RVC voice-conversion engine, including listing and selecting models, setting pitch and index ratio, starting or pausing playback, and toggling original-voice bypass.

Provides text-to-speech synthesis using Microsoft Edge's free TTS engine, supporting multiple voices, languages, and audio output options (base64 or file).

Provides local vision and audio perception for MCP-compatible agents, enabling them to read images, transcribe text from visual media, analyze videos, and convert speech to text entirely on-device. It is privacy-focused with no cloud upload or API keys by default, using Ollama for inference.

Transcribes meeting recordings with speaker diarization and automatic speaker identification, leveraging project context to generate minutes. Supports various ASR engines and offers privacy-preserving local processing.

Enables MCP clients to submit completed speech or typed text to a self-hosted decision engine and receive judgments such as ordinary speech, command, retained thought, uncertainty, or no action, along with session and decision inspection.

A text-to-speech MCP server that enables AI assistants to speak using the VOICEVOX engine with support for multi-character conversations. It features queue management, low-latency streaming via FFplay, and cross-platform playback across Windows, macOS, and Linux.

Provides VOICEVOX text-to-speech as an MCP tool. Requires a running VOICEVOX engine on localhost.

Local speech recognition MCP server powered by FunASR and SenseVoice. It provides privacy-friendly audio transcription through a transcribe_audio tool.

Provides read-only access to local voice-to-text transcription history from the SayAll wireless mic app on macOS. Enables MCP-compatible assistants to query and paginate past transcripts by time and app with strong privacy boundaries.

Enables ChatGPT on the web to generate and play spoken audio through an embedded player, exposing a speak tool that renders text with selectable styles and scenes via configurable text-to-speech providers. It routes each request across ElevenLabs, Cartesia, Azure Speech, and DashScope in a configured failover order, picking providers by language and scenario.

Local Korean text-to-speech MCP server using a fine-tuned CosyVoice2 model, enabling voice synthesis directly from Claude or Codex with privacy and no API costs.

An MCP server that audits voice-assistant tools for prompt injection, tool poisoning, privacy leaks and speech-unfriendly answers. It exposes scanning, PII/link checks, voice rewriting and server-audit tools so users can test whether MCP tools are safe and voice-ready.