WellSaid Claude Connector
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PORT | No | Port for HTTP transport | 8787 |
| MCP_TRANSPORT | No | Transport type: stdio or http | stdio |
| WELLSAID_API_KEY | Yes | Your WellSaid API key | |
| WELLSAID_API_BASE | No | Override the API base URL | https://api.wellsaidlabs.com/v1 |
| WELLSAID_OUTPUT_DIR | No | Where audio/caption files are saved | ./output |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| wellsaid_list_voicesA | List the available WellSaid voice avatars. Returns each voice's speaker_id (the value you pass to wellsaid_generate_speech) plus name, style, accent, gender, and language when the API provides them. Optionally filter with a case-insensitive substring search. |
| wellsaid_list_voice_criteriaA | Summarize the distinct styles, accents, genders, languages, and characteristics across all voices — useful for narrowing a wellsaid_list_voices search. |
| wellsaid_generate_speechA | Convert text to speech with a WellSaid voice and save it to disk. Synchronous. Text is limited to 1000 characters (any AI Director markup counts). Optional prosody (pitch/tempo/loudness) wraps the text in AI Director tags and uses the 'caruso' model — when you pass these, the text is treated as literal (special characters escaped), so don't also hand-author inline tags in the same call. For hand-authored inline tags (e.g. ), leave pitch/tempo/loudness unset. Get a speaker_id from wellsaid_list_voices first. |
| wellsaid_generate_speech_with_timingA | Render text to speech and also get word-level timing and caption files. The API returns a zip (audio + JSON timing + SRT + VTT); this tool extracts and saves each file. Supports the same voice/model/prosody options as wellsaid_generate_speech. |
| wellsaid_create_clip_asyncA | Start an asynchronous TTS render and return the clip id/status. Use for long or batched jobs, then poll with wellsaid_get_clip. Returns the raw job record from the API. |
| wellsaid_get_clipA | Fetch the status/metadata of a clip by id. If download is true and the clip is ready, save its audio to disk (when the API returns a URL or audio payload). |
| wellsaid_list_clipsA | List recent clips created on the account (id, status, metadata). |
| wellsaid_combine_clipsA | Merge multiple existing clips into a single audio file with optional pauses. Provide clip ids in order. pause_durations is a single number applied between all clips, or an array of length (clip_ids.length - 1) for per-gap control. |
| wellsaid_get_respelling_suggestionsA | Get WellSaid respelling suggestions for a word. Use the returned value in an inline AI Director tag, e.g. pizza, to control pronunciation. |
| wellsaid_requestA | Escape hatch for endpoints without a dedicated tool (e.g. replacement libraries). Makes an authenticated request to any path under the API base. JSON responses return as text; binary responses are saved to disk when save_as is provided. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 10 tools
Most tools have clearly distinct purposes (listing voices, summarizing criteria, generating speech, managing clips, combining clips, respelling suggestions, and a request escape hatch). However, the three generation tools—generatespeech, generatespeech_with_timing, and createclip_async—overlap in function, though descriptions clearly distinguish sync vs async and timing output.
All tool names follow a consistent wellsaid_verb_noun snake_case pattern (e.g., wellsaid_list_voices, wellsaid_generate_speech, wellsaid_combine_clips). No mixing of styles or inconsistent verb choices.
10 tools is well within the ideal range, covering the full TTS workflow—voice discovery, generation (sync/async/timing), clip management, combination, and respelling—without unnecessary bloat.
The tool surface is comprehensive for TTS generation and clip management, including sync/async rendering, timing/captions, and clip combination. Minor gaps like no dedicated update/delete for clips are mitigated by the wellsaid_request escape hatch, but the core lifecycle lacks these operations.