"Automating Testing and Debugging for LLMs in Production" matching MCP servers:

Enables an Alexa+ host to place and manage customer-care calls by building a case, navigating phone menus, waiting on hold, explaining the issue, guarding sensitive disclosures, handing off identity checks, recording ticket numbers and promises, and scheduling follow-ups.

A local stdio MCP server that enables validated access to official ElevenLabs APIs with cost controls, local media storage, and audio export for Godot.

Enables AI agents to synthesize and play short summaries aloud while keeping full answers on screen, with queued speech and stop control using online natural voices or offline synthesis.

Enables voice-driven family check-ins and safety escalation through five MCP tools over Streamable HTTP, recording check-ins, scheduling recurring reminders with grace windows, detecting missed check-ins, raising critical help alerts, and resolving them from a live dashboard. All reads and writes flow through these tools into a SQLite store, with safety rules ensuring explicit help requests override upbeat wording and never duplicate or downgrade alerts.

Enables real-time full-duplex voice agents to classify user interruptions, backchannels, and acoustic echo, and to arbitrate turn-taking with context rollback via MCP.

Enables real-time voice agents to arbitrate full-duplex speech by classifying user interruptions, backchannels, and acoustic echo while supporting context rollback.

Enables an AI assistant to place user-approved phone calls in the callee's language through your own Bland or Twilio account, with language detection, script generation, cost estimates, and safety guard rails. Includes a preview-then-confirm gate, plus call status, transcripts, and summaries.

Enables AI clients to watch and hear videos from local files or YouTube, Vimeo, and Loom links by returning timestamped key frames, a transcript, a contact sheet, and a combined timeline, with all processing done locally.

Transcribe audio/video URLs and detect silence for cutting, with pay-per-call billing in USDC via x402, no account or API key required.

Enables MCP-capable assistants to transcribe local audio files or URLs and perform speaker diarization for Spanish and Portuguese audio, with options for speaker count hints, domain prompts, and transcript retrieval in multiple formats.

Fast MCP Telegram — Production-grade Telegram MCP server with direct MTProto API, multi-user ACL, dual transport (stdio + HTTP SSE), voice transcription, and context-optimized design. One-command setup with uvx fast-mcp-telegram

Official MCP server for Theta EdgeCloud's On-Demand Model APIs, providing access to 20+ AI models including image generation, audio transcription, and LLMs, directly from MCP-compatible clients.

Enables AI assistants to transcribe links (YouTube, Google Drive, Dropbox), list and search your recordings for spoken words and moments, and read transcripts in json/txt/srt/vtt, summaries and metadata. It also lets them request new summaries and delete recordings, all through your own account with API-key or OAuth sign-in.

Turns text into local audio files (MP3) via OpenRouter speech models, enabling AI coding agents to generate voiceover narrations for video production.

Enables Claude Code to leverage Chinese LLMs for multimodal tasks including image analysis, audio transcription, and deep synthesis via MCP protocol.

Provides local audio analysis tools for LLMs, enabling transcription, conversation dynamics, prosody analysis, and visual inspection without API keys.

Enables LLMs to provide voice status updates via OpenAI TTS, ideal for background task progress reports.

An MCP server that enables LLMs to generate spoken audio from text using OpenAI's Text-to-Speech API, supporting various voices, models, and audio formats.

Enables MCP-compatible LLMs to control Amazon Alexa devices, including announcements, voice commands, smart-home group management, routines, media playback, and sensor state queries.

Provides intelligent transcript processing capabilities for Claude, featuring natural formatting, contextual repair, and smart summarization powered by Deep Thinking LLMs.

An MCP server that enables LLMs to access the NijiVoice API for text-to-speech generation, supporting features like fetching available voice actors and checking credit balance.

An MCP (Model Context Protocol) server that provides seamless integration between Fish Audio's Text-to-Speech API and LLMs like Claude, enabling natural language-driven speech synthesis.

Transcribe YouTube videos for LLM chat applications. Supports fetching transcripts in multiple languages and includes an ad-removal instruction.

Enables coding agents to build, test, and deploy production voice agents with real phone numbers directly from the editor, including live call monitoring and sandboxed development.

Enables any AI to place real phone calls in Spain and Mexico with an AI voice agent, then retrieve call status, summaries, collected data, and transcripts.

Production-ready MCP server for the deAPI REST API, exposing AI capabilities like audio transcription, image generation, OCR, video generation, and text-to-speech as MCP tools.

Enables MCP clients to place real phone calls to businesses in the US and Canada, navigating phone menus, waiting on hold, asking the agent mid-call for information, and returning structured answers with transcripts. Supports booking, quotes, cancellations, and multi-business calls in parallel.

Lets AI assistants manage voice agents, calls, and phone numbers, exposing read and write tools plus spend tools gated behind explicit opt-in.

Provides adaptive jitter buffering and packet loss concealment for real-time voice streams, ensuring smooth audio playout in unstable network conditions.

Provides voice summaries after each AI request in Cursor, Cline, or any MCP-supported editor, allowing users to hear what was done and stay informed without staring at the screen.

Gives any MCP client access to your WhatsApp, locally on macOS: full-text search across years of chats, photos, PDFs and Word documents, and voice notes transcribed on-device with Whisper. It can also draft messages, or send them through the official app with built-in safeguards.

Enables assistants to open scripts in a centered, multilingual reader with direct paste, locale-aware word tracking, manual sentence navigation, pause/auto-scroll, and optional browser or OpenAI speech recognition.

A self-contained daemon that watches a directory, transcribes anything that lands in it, and presents the results to any AI agent through a standard protocol — no cloud, no subscriptions, no vendor lock-in.

Enables agents to transcribe videos in their original language, review uncertain speech against screenshots, and submit corrections while preserving the original transcript and generating exports.

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.

Enables deterministic psychoacoustic modeling of dichotic hearing thresholds, interaural phase differences, brainstem processing scores, and cocktail-party speech-in-noise intelligibility through JSON-RPC tools.

Enables LLMs to transcribe, analyze, and translate audio/video content through Gladia's API.

Enables AI agents to ring or text a user's actual iPhone, read a question or update aloud, and return the spoken answer back to the agent as text. Calls and messages land in a single titled thread per agent, with tools for polling results, waiting for replies, and labeling conversations.

Enables AI agents to speak and listen in real-time with interruption handling, using local ML models and hot-swappable adapters.

Enables hands-free voice collaboration with a coding agent: you speak requirements aloud and hear only short, plain-language progress and completion summaries while the agent edits code, runs tests, and fixes bugs in the background. It also handles continuous speech recognition with automatic turn detection, barge-in interruption, and spoken approval prompts before risky operations.