MCP server category
Voice and conversation MCP servers: comparison and full list
The Voice and conversation category lists 13 open-source MCP servers, checked on GitHub. The best scored on our grid are Voicemode, AgentPhone, and PlaceCall. 62% received a commit in the last three months.
What Voice and conversation MCP servers are for
Voice and conversation MCP servers connect an AI assistant to speech recognition, speech synthesis or telephone services. Depending on the integration, you can speak to an agent, listen to written material or ask it to handle a telephone exchange. These tools serve people who prefer spoken interaction, support teams and developers building voice interfaces. Voicemode supports voice conversations through speech recognition and synthesis. Kokoro tts MCP turns text into audio files. A practical request might be: “Read this project update aloud so I can check whether the wording is clear while I walk.”
Example uses:
- Listen to a written report to check its wording, pacing and overall clarity before sharing.
- Turn a spoken request into text that your assistant can process within its usual workflow.
- Generate an audio file from approved text for a presentation, demonstration or recorded announcement.
- Use PlaceCall to ask a US business for a quote and review the returned transcript.
The category in figures
Of the 13 servers in the category (0% of the directory), 3 are official vendor implementations, 92% declare a licence, 5 offer a remote server and 5 need at least one API key or token according to the MCP registry. The median repository has 14 stars.
All Voice and conversation MCP servers
Click a header to sort. Score out of 100: adoption, maintenance and framework (licence, registry, official status); risk: system access, secrets, licence, abandonment.
| Server | Stars | Last commit | Licence | Risk | Score |
|---|---|---|---|---|---|
| Voicemode Natural voice conversations for AI assistants - STT/TTS via MCP | 1,417 | Oct 21, 2026 | MIT | moderate | 83 |
| AgentPhone Give AI agents real phone numbers, messages, and voice calls via MCP. | 124 | Oct 22, 2026 | MIT | moderate | 66 |
| PlaceCall Real phone calls to US businesses: book, ask, get quotes. Returns outcome and transcript. | 105 | Oct 23, 2026 | MIT | low | 65 |
| Kokoro tts MCP Text-to-speech with the open-weight Kokoro TTS models, saving MP3s locally or uploading them to an S3 bucket. | 83 | Sep 12, 2025 | Apache-2.0 | moderate | 36 |
| Loudkit Local text-to-speech from an on-device model with multiple voices and languages, cloned voice profiles and WAV, FLAC, MP3 or Opus output. | 26 | Oct 20, 2026 | Apache-2.0 | low | 59 |
| Claude code tts Text-to-speech for Claude Code using OpenAI's TTS API, with multiple voices and cross-platform playback. | 21 | Jan 25, 2026 | MIT | low | 38 |
| Daisys MCP Generate text-to-speech and text-to-voice audio with the DAISYS platform, and play and store the generated audio. | 14 | Jul 31, 2025 | — | high | 18 |
| MCP transcribe Transcribe audio and video files and voice memos and work with their text content. | 8 | Apr 3, 2026 | MIT | moderate | 39 |
| Perspective AI ◆ Design Perspective AI conversation agents that replace static forms, analyze conversations, deploy embeds and automate follow-ups. | 6 | Sep 1, 2026 | MIT | low | 63 |
| Babelscribe Local speech-to-text and SRT/VTT subtitles for audio and video in 99 languages on your own AMD, NVIDIA, Intel or Apple GPU (whisper.cpp via | 6 | Oct 10, 2026 | MIT | low | 58 |
| Spoken Search podcast episodes and fetch published transcripts as Markdown with real speaker names via the Spoken API. | 5 | Oct 6, 2026 | MIT | moderate | 57 |
| Zihin ◆ Build and operate AI agents on the Zihin.ai platform: manage agents, schemas, triggers, budgets and approvals, run diagnostics and chat with | 0 | Oct 5, 2026 | MIT | moderate | 60 |
| Lobbyvoices MCP ◆ Tools from Lobby, an AI receptionist for small businesses: write phone scripts and IVR menus, generate ElevenLabs agent prompts and role-pla | 0 | Jul 3, 2026 | MIT | low | 43 |
Official servers
Published by the service's own vendor (per the awesome-mcp-servers list): they usually track the API closely and are better maintained.
- Perspective AI 6 ★
- Zihin 0 ★
- Lobbyvoices MCP 0 ★
Choosing and securing
Start by separating speech synthesis, transcription and telephone requirements. Compare supported languages, audio formats and dependencies on external providers. Local models may reduce transfers but require suitable computing resources. Review repository maintenance, licensing, voice usage rights and secret handling. For telephone integrations, look for approval controls, accessible activity records and clear documentation of supported destinations and charges before connecting an account.
Audio can expose someone's identity, telephone number or confidential information. Check where recordings and transcripts are processed, stored and shared. A telephone call can disclose information or create commitments on your behalf. Require approval of the recipient and purpose before calling. Treat speech received from other people as external content, including any instructions that attempt to redirect the assistant or trigger further actions.
Before connecting a server to an agent with access to your data, run the AI agent security check.
Most recent
- Babelscribe created October 6, 2026
- Loudkit created August 22, 2026
- PlaceCall created July 24, 2026
- Lobbyvoices MCP created June 9, 2026
- Spoken created June 4, 2026
Other categories
Learn more
Frequently asked questions
What is the best Voice and conversation MCP server?
On our grid (adoption, maintenance, framework), Voicemode (83/100), AgentPhone (66/100), and PlaceCall (65/100) come first. The right choice mostly depends on the service you need and the rights you are willing to grant.
Are these servers free?
92% declare an open-source licence and are free to use. 5 need an API key: the connected service may be paid.
Can I generate speech without sending my text to a remote provider?
Some integrations support local processing. Loudkit describes speech synthesis using an on-device model. Check the complete setup, including logs, downloads and any supporting services. Saving the resulting audio on your computer does not, by itself, establish that the text was processed locally throughout the workflow.
Can a voice server make telephone calls for me?
Certain servers provide telephone functions, including AgentPhone and PlaceCall within their documented scope. A speech synthesis server alone does not provide calling. Before using one, check supported destinations, charges, information disclosure and whether consequential requests can be reviewed before the agent communicates them to another person.
Related articles
- Build your own MCP server in 30 minutesOctober 24, 2026
- Automate your inbox with an AI agentOctober 18, 2026
- MCP, A2A, and tool calling: agent protocols explainedOctober 17, 2026
- Build an AI agent with n8n, without codingOctober 16, 2026
Neutrality. No affiliate links. No vendor pays to be reviewed, rated or ranked. Reviews are directed by François Chollet, neutral and factual, and every data point is dated and sourced. How we score