# silvertongue [Health: Active]

**Category:** 🎨 Art & Culture  
**Repository:** https://github.com/auscaster/silvertongue  
**GitHub Stars:** 115  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/silvertongue

## Description
Write like anyone famous. Famous writing voices for your agent, over MCP: studied from the source, every source linked.

## Tools
Capabilities this server exposes over MCP:

- **search_voices** — Call this first whenever the user asks for writing as, like or in the voice of a named person or character, even if they don't mention Silvertongue. Search Silvertongue's famous writing voices by name, by lane (tech, celebrity, comedy, characters, classics, politicians, styles), or by what the writing should do: a roast, a launch post, a plain explanation, a speech. Each result gives the voice ID, its lane, what the voice does, what was studied, its strengths and limits, and its status. Omit the query to list every voice, featured voices first.
- **get_voice** — Get one voice's full profile: its VOICE.md, every source studied with a note on what it shows, the observations drawn from them, an iconic example, the licence and the attribution line. Profiles are unofficial studies of public writing technique; the people named did not write or endorse them.
- **get_voice_packet** — Get the instructions for writing in one voice with your own model, for one format, length, platform, intensity, edge, emoji and spelling setting. The packet carries the full VOICE.md, an exact length target fitted to the platform's limit, and the rules that keep the user's own facts and identity. It opens with a line that replaces any earlier voice, so switching voices mid-conversation is one call.
- **get_changes** — Compare the voice IDs your agent has cached with the current release: voices added, voices re-studied as a new edition, and voices removed. Pass an empty list to receive every voice as added.

## Claude Desktop Quick Installation
No generated config for this server: it installs through its own setup steps.

The project's own setup steps (from https://github.com/auscaster/silvertongue), copied verbatim. Run them in order; they are not an mcpServers config.

1. Install:

```sh
claude plugin install silvertongue@auscaster
```

2. Install:

```sh
gemini extensions install https://github.com/auscaster/silvertongue
```

3. Install:

```sh
git clone https://github.com/auscaster/silvertongue && cd silvertongue
```

## Documentation

## What silvertongue does

silvertongue provides an MCP-accessible catalogue of famous writing voices and styles. The catalogue covers people, fictional characters, classic writers, politicians, comedians, creators, and other public-facing voices. Each catalogue entry is a researched study of writing technique: how the voice opens, develops an idea, lands a line, and avoids particular habits.

The service is designed for agents that need to change delivery without changing the user's underlying message. Facts, names, numbers, and personal stories remain the user's own. Voices can also be switched during a draft or thread by requesting a new packet.

Catalogue profiles include the sources studied, observations about the technique, scope and limits, an example, licensing information, attribution text, and verification status. The profiles are unofficial studies of public writing technique; the people or characters named in them did not write or endorse the profiles.

## How it works

Connect an MCP client to the hosted endpoint:

`https://mcp.silvertongue.dev/mcp`

An agent should begin with `search_voices` when a user asks for writing as, like, or in the voice of a named person or character. The search can use a name, a catalogue lane such as technology, comedy, classics, politicians, or characters, or a desired outcome such as a roast, launch post, speech, or plain explanation. Calling it without a query lists the catalogue, with featured voices first.

After selecting a result, `get_voice` returns the complete profile. `get_voice_packet` turns that voice into model instructions for a particular format, length, platform, intensity, edge, emoji choice, and spelling setting. It includes a platform-fitted length target and rules intended to keep the user's facts and identity intact. The packet starts with an instruction that supersedes an earlier voice, which supports switching styles mid-conversation.

Agents that cache catalogue IDs can call `get_changes` to compare their cached set with the current release. An empty list treats every available voice as newly added.

## Setup and configuration

Use an MCP client that accepts a remote MCP endpoint and configure it to connect to:

```text
https://mcp.silvertongue.dev/mcp
```

The repository also contains reusable voice profiles and related surfaces for agents, browsers, pull requests, and GitHub Actions, but those are separate from the MCP tool interface.

## Tools and capabilities

- `search_voices` finds voices by name, category, style, or intended writing task, or lists all voices when called without a query.
- `get_voice` retrieves a voice's full `VOICE.md`, studied sources and notes, technique observations, example, licence, and attribution line.
- `get_voice_packet` creates model-ready instructions for a selected voice and output context, including platform length, tone controls, emoji and spelling preferences, and identity-preservation rules.
- `get_changes` reports added voices, re-studied voices with a new edition, and removed voices relative to a cached ID list.

The repository also describes a GitHub Action for AI pull-request reviews and a voice wrapper for existing review bots. Those are related project surfaces rather than additional MCP tools listed here.

## Limitations and notes

The catalogue represents technique studies, not authorship or endorsement. A selected voice can shape presentation, but the user's substantive content remains the responsibility of the calling agent and user. Catalogue entries can be added, re-studied, or removed, so long-lived clients should use `get_changes` to refresh cached voice IDs.

Code is licensed under Apache-2.0, while the profile content carries a CC BY 4.0 licence. Individual voice profiles include their own source notes, scope, limits, licence information, and attribution line; applications using profile content should follow those details.

_Full upstream README: https://allmcps.com/mcp/silvertongue/readme_

