This server connects MacWhisper (a local macOS transcription app) to Claude Desktop, enabling on-device audio transcription and management — all without sending data to the cloud. Supported formats: .m4a, .mp3, .mp4, .mov, .wav, .aiff, .flac.
Transcribe audio (transcribe_audio): Convert local audio files to text, with optional model selection and the ability to save results to MacWhisper's history.
List installed models (list_models): View all transcription models installed in MacWhisper, with the active model clearly marked. Returned model IDs can be passed directly to transcribe_audio.
Cancel a transcription (cancel_transcription): Stop a transcription currently in progress.
List allowed paths (list_allowed_paths): Check which directories the server is permitted to read audio files from.
Watch a folder (start_watch): Monitor a folder and automatically transcribe new audio files dropped into it, moving completed files to a done/ subdirectory.
Stop the folder watcher (stop_watch): Halt the active folder-watching process.
Get watch results (get_watch_results): Fetch transcription results from the folder watcher (file name, transcript, destination, errors) and clear the queue.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@macwhisper-mcp-server Transcribe ~/Desktop/meeting.m4a and summarize key points"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Local MCP server that connects MacWhisper to Claude Desktop.
What it does: Drop an audio file on your Desktop, then ask Claude to transcribe it, summarise it, or pull out action items — in one step. MacWhisper does the transcription on your Mac; Claude does the thinking. Nothing leaves your machine. No cloud APIs. No data ever leaves your Mac.
Audio file → MacWhisper CLI → MCP server → Claude Desktop
macOS (MacWhisper is macOS-only)
MacWhisper — installed and licensed
MacWhisper CLI enabled: open MacWhisper → Settings → Advanced → Command-Line Tool → Install. This places mw at /usr/local/bin/mw.
Python 3.13.x via pyenv
Installing MacWhisper via Homebrew:
brew install --cask macwhisperAfter installation, enable the CLI in MacWhisper Settings as above. When you later run brew upgrade --cask macwhisper, the CLI symlink updates automatically — no re-install needed.
Related MCP server: Plaud Notes MCP Server
brew tap docdyhr/tap
brew install docdyhr/tap/macwhisper-mcp-serverThis installs the macwhisper-mcp binary into your Homebrew prefix. Upgrade later with brew upgrade docdyhr/tap/macwhisper-mcp-server.
pip install macwhisper-mcp-serverOr from source:
git clone https://github.com/docdyhr/macwhisper-mcp-server.git
cd macwhisper-mcp-server
pyenv install 3.13.13 # skip if already installed
pyenv local 3.13.13
python -m venv .venv
source .venv/bin/activate
pip install -e .Verify the MacWhisper CLI is reachable:
mw versionEdit ~/Library/Application Support/Claude/claude_desktop_config.json and add:
{
"mcpServers": {
"macwhisper": {
"command": "macwhisper-mcp",
"args": [],
"env": {
"MACWHISPER_ALLOWED_PATHS": "~/Desktop:~/Downloads",
"FASTMCP_CHECK_FOR_UPDATES": "off"
}
}
}
}Restart Claude Desktop.
Note: Audio files must be saved to your Mac's filesystem (Desktop, Downloads, or another allow-listed folder) before asking Claude to transcribe them. Files uploaded directly to the Claude chat window live in Claude's container and are not accessible to the local MacWhisper CLI.
In Claude Desktop, ask:
Transcribe ~/Desktop/memo.m4a
You should see a transcribe_audio tool call appear, followed by the transcript.
Tool | Description |
| Transcribe an audio file and return the transcript as plain text. |
| List transcription models installed in MacWhisper, plus whisper-cpp models if |
| Cancel the currently running transcription |
| Return the directories the server is allowed to read from |
| Watch a folder and auto-transcribe new audio files into |
| Stop the active folder watcher |
| Return completed watch-folder transcriptions and clear the queue |
Supported audio formats: .m4a .mp3 .mp4 .mov .wav .aiff .flac
All configuration is via environment variables. Pass them through the env dict in claude_desktop_config.json (for Claude Desktop) or set them in .env for local development.
Env var | Default | Description |
|
| Colon-separated list of directories the server may read from |
| auto-detected | Path to the |
|
| Log file path (never stdout — that's reserved for MCP) |
| none | Colon-separated |
|
| Path to the |
| none | Directory containing your GGML |
Local development: copy .env.example to .env and adjust. With direnv, .envrc exports .env automatically. Without direnv: source .env.
If you regularly transcribe recordings in a specific language, map a subfolder to
it instead of passing language on every call:
"MACWHISPER_LANGUAGE_DEFAULTS": "~/Desktop/DK=da:~/Desktop/DE=de"Drop a file in ~/Desktop/DK/ and transcribe_audio passes --language da
automatically. An explicit language argument on the tool call always wins over
the directory default.
transcribe_audio(..., engine="whisper-cpp") transcribes using a standalone
whisper.cpp binary instead of
MacWhisper — useful if you don't have a MacWhisper license, or want a fully
open-source local path. It does not touch MacWhisper in any way.
Setup:
brew install whisper-cppHomebrew installs the whisper-cli binary only — no models. Download a GGML
model yourself (this server never downloads anything over the network) from
huggingface.co/ggerganov/whisper.cpp,
e.g. ggml-base.en.bin, into a directory of your choice, then point the server at it:
"MACWHISPER_WHISPERCPP_MODEL_DIR": "~/whisper-models"Then call the tool with the model's filename (not a MacWhisper engine:model-id
string):
Transcribe ~/Desktop/memo.wav using the whisper-cpp engine with model ggml-base.en.bin
Limitations (v1):
Input formats: .wav, .mp3, .flac only — not .m4a/.mp4/.mov/.aiff. This
is whisper.cpp's own native format support; convert other formats first (e.g. with
ffmpeg) or use the default MacWhisper engine, which handles all supported formats.
persist=true is not supported — whisper.cpp has no history mechanism.
Default language is English (en) if neither language nor a directory default
is set — unlike MacWhisper, which defers to the app's own language selection.
source .venv/bin/activate
pip install -e ".[dev]"
# Tests
pytest -q
# Lint + format
ruff check .
ruff format .
# Pre-commit hooks (one-time setup)
pip install pre-commit
pre-commit install
# Smoke-test against a real audio file (server must not be running in Claude Desktop)
python scripts/smoke_test.py ~/Downloads/Test.m4atail -f ~/Library/Logs/macwhisper-mcp.logAll file paths are resolved (symlinks followed) and checked against the MACWHISPER_ALLOWED_PATHS allow-list before anything reaches the CLI.
subprocess.run is always called with an argv list — never shell=True.
No network calls. Ever.
See PRD §7 for the full threat model.
Uploaded files: Files dragged into the Claude chat window live in Claude's container and are not accessible to the local MacWhisper CLI. Save the file to your Desktop or Downloads folder (or another allow-listed directory), then ask Claude to transcribe it from there.
Danish letter names: Whisper may phonetically approximate letter names (e.g. "Æ, Ø, Å" → "E, Y, U") when they are spoken in isolation. Letters inside words transcribe correctly. This is a Whisper engine limitation, not a bug in this wrapper. See PRD §12.
Cold-start latency: First transcription after MacWhisper launches takes ~13s (model load). Subsequent calls are ~2s.
MIT — see LICENSE.
Cancel the currently running transcription, if any.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, and the description does not disclose behavior when no transcription is running (e.g., error vs no-op). Without this, the agent lacks full understanding of edge cases.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single sentence, no wasted words. Description is appropriately sized and front-loaded with the action.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no parameters and an output schema (not shown), the description is minimal. It fails to specify the effect when no transcription exists, but with output schema, return values may be documented elsewhere. Adequate but not thorough.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters exist, so schema coverage is 100%. Per guidelines, 0 parameters yields baseline 4. Description adds no additional parameter info, which is fine.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool cancels the currently running transcription, which is a specific verb+resource. It distinguishes from siblings like transcribe_audio (starts) and stop_watch (stops watch, not transcription).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Implies use when a transcription is running, but does not explicitly state when not to use or what happens if none is running. However, the context is clear enough for a simple action.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return completed watch-folder transcriptions and clear the queue.
Each entry contains: file, transcript, destination, error.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description discloses the important behavioral trait of clearing the queue after returning results, and lists the fields in each entry. It does not mention idempotency or behavior when queue is empty.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two sentences, front-loaded main action, no extraneous information. Every sentence earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the presence of an output schema, the description provides helpful detail about return fields. However, it lacks information about pagination or behavior when no results exist.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are no parameters, so schema coverage is 100%. The description adds no parameter info, which is appropriate. Baseline score for zero parameters is 4.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool returns completed watch-folder transcriptions and clears the queue, distinguishing it from sibling tools like 'start_watch' and 'transcribe_audio'.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage after starting a watch and after transcriptions complete, but does not explicitly state when to use this tool versus alternatives or when not to use it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return the directories this server is allowed to read audio from.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description fails to disclose behavioral traits like authentication needs or side effects, providing minimal transparency for a read operation.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single concise sentence with no unnecessary words, front-loaded with the verb and resource.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The tool is simple with no parameters and an output schema, so the description is nearly complete; however, it could mention that it returns a list of paths.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has no parameters, so the description adds no param info; but with 100% schema coverage and zero parameters, a baseline score of 4 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool returns directories the server is allowed to read audio from, with a specific verb and resource, distinguishing it from siblings like transcribe_audio.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage for verifying allowed paths before audio operations, but lacks explicit when-to-use or alternatives, making it adequate but not fully guided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return the transcription models available across all engines.
MacWhisper entries are formatted as engine:model-id — Display Name [active]
where [active] marks the model currently selected in MacWhisper; pass the
engine:model-id string as model with the default engine="macwhisper".
whisper-cpp entries (if MACWHISPER_WHISPERCPP_MODEL_DIR is configured) are
formatted as filename [whisper-cpp]; pass the filename as model with
engine="whisper-cpp". Requires the MacWhisper CLI to be reachable even if
you only intend to use the whisper-cpp engine.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the burden of disclosing behavior. It reveals the output format for both MacWhisper and whisper-cpp entries, the conditional presence of whisper-cpp models based on an environment variable, and the requirement that the CLI be reachable. This is substantial behavioral disclosure, though it doesn't cover error scenarios or exact return structure beyond the string formats.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is front-loaded with the purpose and then provides necessary formatting details in a structured way. Each sentence adds value for the consumer, though the second and third sentences are somewhat dense. It is not overly verbose for the information it conveys.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has an output schema (not shown), the description explains the semantic meaning of the output strings (active marker, engine prefixes) and how to feed them into transcribe_audio. It also covers prerequisites and conditional behavior, making it complete for a list tool with no params.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, so the input schema is empty and schema coverage is 100%. The description correctly adds no parameter-specific details. Per the rubric, 0 params warrants a baseline of 4, and there is nothing to add beyond that.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The opening sentence 'Return the transcription models available across all engines' uses a specific verb ('Return') and resource ('transcription models') with a clear scope. This clearly distinguishes it from sibling tools like transcribe_audio or start_watch, which perform different actions.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explains when and how to use the tool: it lists models formatted for use with transcribe_audio, and it notes the prerequisite that the MacWhisper CLI must be reachable even for whisper-cpp. It does not explicitly name alternatives or exclusions, but the context is clear enough for selecting this tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Start watching a folder for new audio files to auto-transcribe.
New audio files dropped into folder are transcribed automatically
and moved to a "done" directory. By default this is <folder>/../done;
override it with the MACWHISPER_WATCH_DONE_DIR env var. Both the
incoming folder and the done directory must be inside the configured
allow-list, otherwise the call is rejected. Call get_watch_results()
to retrieve completed transcriptions.
| Name | Required | Description | Default |
|---|---|---|---|
| folder | Yes | Absolute or ``~``-prefixed path to the incoming directory. |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It explains that files are automatically transcribed and moved to a 'done' directory, mentions the env var override, and states the allow-list requirement with rejection behavior. It also points to a sibling for results, giving a comprehensive view of the tool's side effects and constraints.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is tightly structured with a one-line purpose, followed by two sentences of essential operational details and a pointer to the results tool. Every sentence contributes unique information without redundancy or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (background watch, side effects, allow-list constraint) and the presence of an output schema, the description covers all necessary aspects: what it does, how it behaves, what constraints apply, and how to retrieve results. It equips an agent to decide when and how to use it effectively.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already describes the 'folder' parameter as an absolute or ~-prefixed path. The description adds meaningful context: it is the incoming directory for watched files, and both this directory and the derived done directory must be inside the allow-list. This goes beyond the schema by clarifying parameter role and constraints.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with 'Start watching a folder for new audio files to auto-transcribe,' which is a specific verb-object phrase that clearly states the tool's function. It is easily distinguishable from sibling tools like transcribe_audio, stop_watch, and get_watch_results, which involve different actions.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly implies when to use this tool: when you want automatic transcription of newly dropped files. It also guides the user to call get_watch_results() to retrieve transcriptions. However, it does not explicitly contrast this with transcribe_audio for one-off transcription, so the exclusion guidance is not fully spelled out.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Stop the active folder watcher.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must convey behavioral traits, but it only states the basic action. It omits important details such as side effects (e.g., whether watched files are released), error conditions, or required permissions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, complete sentence with no unnecessary words. It is front-loaded and efficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
While the tool is simple (zero parameters), the description does not mention the output or any side effects, despite an output schema being present. It is minimally adequate but leaves some context uncovered.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The tool has zero parameters, so the description does not need to add parameter meaning. The schema already covers all 0 parameters, meeting the baseline for high schema coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Stop') and the resource ('the active folder watcher'), making the purpose unambiguous. It effectively distinguishes from siblings such as 'start_watch'.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives (e.g., 'cancel_transcription' or 'start_watch'). It lacks context on prerequisites or expected state.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Transcribe a local audio file and return the transcript.
IMPORTANT: path must be a file on the user's Mac filesystem inside the
configured allow-list (typically ~/Desktop or ~/Downloads). Files uploaded
to the Claude chat window are NOT accessible — ask the user to save the file
to their Desktop or Downloads folder first.
If this tool returns an access-denied error, do NOT attempt to transcribe the file by any other means (e.g. downloading a model, calling an external API, or using in-process speech recognition). Simply tell the user to save the file locally and retry with this tool.
| Name | Required | Description | Default |
|---|---|---|---|
| path | Yes | Absolute or ``~``-prefixed path to an audio file on the local Mac filesystem inside the configured allow-list. Supported formats: m4a, mp3, mp4, mov, wav, aiff, flac (whisper-cpp engine only: wav, mp3, flac). | |
| model | No | Optional model override. For the default ``engine="macwhisper"``, an engine:model-id string, e.g. "whisperkit:openai_whisper-large-v3-v20240930" — use ``list_models()`` to see what is installed. For ``engine="whisper-cpp"``, this is REQUIRED and must be the filename of a GGML model inside ``MACWHISPER_WHISPERCPP_MODEL_DIR`` (not an engine:model-id string). | |
| engine | No | "macwhisper" (default) — routes through the MacWhisper CLI, or "whisper-cpp" — an independent local backend that does not use MacWhisper at all. Requires whisper-cpp installed (`brew install whisper-cpp`) and MACWHISPER_WHISPERCPP_MODEL_DIR set. | macwhisper |
| persist | No | If True, save the transcription to MacWhisper's history. Defaults to False. Not supported by the whisper-cpp engine. | |
| language | No | Optional ISO 639-1 code (e.g. "da", "de") or "auto" to force the transcription language. If omitted, a per-directory default is used when the file's folder matches one configured in MACWHISPER_LANGUAGE_DEFAULTS; otherwise the engine's own default applies (MacWhisper: app selection; whisper-cpp: English). |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full transparency burden. It discloses the access restriction (must be inside allowed path), the unavailability of chat uploads, and the recommended behavior on access-denied errors. However, it does not explicitly mention potential side effects like whether persist=True saves to history (covered in schema) or whether the transcription is fully local. It adds substantial context beyond the schema but still has some gaps.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is structurally efficient: a one-sentence purpose, an IMPORTANT constraint block, and a concise error-handling paragraph. Each sentence adds value with no redundancy, and the critical warnings are front-loaded for quick scanning.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (5 parameters, engine alternatives, language defaults) and the presence of an output schema (so return format is documented), the description covers the essential usage context: local file requirement, chat uploads inaccessible, and error handling. It does not explicitly mention when to prefer start_watch for continuous/batch transcription, but the description remains sufficiently complete for an agent to use the tool correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 100% parameter coverage with detailed descriptions for all 5 parameters, so the baseline is 3. The description adds important semantics for the path parameter: it must be on the Mac filesystem in the allow-list, and chat uploads are not accessible. This is not fully captured in the schema's path description, making the overall parameter understanding richer.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a precise verb and resource: 'Transcribe a local audio file and return the transcript.' This clearly identifies the tool's purpose and distinguishes it from sibling tools such as cancel_transcription, list_models, and start_watch, which are support operations rather than the core transcription action.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description explicitly states when to use the tool: for local files inside the allow-list, and it excludes chat uploads. It also provides specific error-handling guidance: if access-denied, do not attempt alternatives (downloading models, external APIs, in-process speech recognition); instead ask the user to save the file locally and retry. This is clear when-to-use and when-not-to-use guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v1.2.0transcribe_audio5 fields changedInput schema / properties / engineAdded value: +{
+ "default": "macwhisper",
+ "description": "\"macwhisper\" (default) — routes through the MacWhisper CLI, or\n\"whisper-cpp\" — an independent local backend that does not use\nMacWhisper at all. Requires whisper-cpp installed\n(`brew install whisper-cpp`) and MACWHISPER_WHISPERCPP_MODEL_DIR set.",
+ "type": "string"
+}Input schema / properties / languageAdded value: +{
+ "anyOf": [
+ {
+ "type": "string"
+ },
+ {
+ "type": "null"
+ }
+ ],
+ "default": null,
+ "description": "Optional ISO 639-1 code (e.g. \"da\", \"de\") or \"auto\" to force\nthe transcription language. If omitted, a per-directory default is\nused when the file's folder matches one configured in\nMACWHISPER_LANGUAGE_DEFAULTS; otherwise the engine's own default\napplies (MacWhisper: app selection; whisper-cpp: English)."
+}Input schema / properties / model / descriptionPrevious value: -"Optional model override in MacWhisper engine:model-id format,\ne.g. \"whisperkit:openai_whisper-large-v3-v20240930\". Use\n``list_models()`` to see what is installed. Defaults to the\nmodel currently selected in MacWhisper."New value: +"Optional model override. For the default ``engine=\"macwhisper\"``,\nan engine:model-id string, e.g.\n\"whisperkit:openai_whisper-large-v3-v20240930\" — use ``list_models()``\nto see what is installed. For ``engine=\"whisper-cpp\"``, this is\nREQUIRED and must be the filename of a GGML model inside\n``MACWHISPER_WHISPERCPP_MODEL_DIR`` (not an engine:model-id string)."Input schema / properties / path / descriptionPrevious value: -"Absolute or ``~``-prefixed path to an audio file on the local Mac\nfilesystem inside the configured allow-list.\nSupported formats: m4a, mp3, mp4, mov, wav, aiff, flac."New value: +"Absolute or ``~``-prefixed path to an audio file on the local Mac\nfilesystem inside the configured allow-list.\nSupported formats: m4a, mp3, mp4, mov, wav, aiff, flac (whisper-cpp\nengine only: wav, mp3, flac)."Input schema / properties / persist / descriptionPrevious value: -"If True, save the transcription to MacWhisper's history.\nDefaults to False."New value: +"If True, save the transcription to MacWhisper's history.\nDefaults to False. Not supported by the whisper-cpp engine."v0.1.0cancel_transcriptionget_watch_resultslist_allowed_pathslist_modelsstart_watchstop_watchtranscribe_audioEach tool has a distinct role: transcription, cancellation, model listing, path listing, and watch lifecycle. The only slight overlap is between transcribe_audio and get_watch_results, both returning transcripts, but they serve different workflows (direct vs. watched).
All tools follow a consistent snake_case verb_noun pattern (e.g., transcribe_audio, start_watch), making the API predictable and easy to navigate.
Seven tools is well within the ideal range; each tool addresses a specific need without redundancy.
The core transcription workflow is covered (transcribe, cancel, list models, list paths), and the watch folder feature adds a complete sub-workflow. Minor gap: no way to retrieve a past transcription result outside of the watch queue, but not essential for the main purpose.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Let ChatGPT, Claude & Cursor use your Mac: email, calendar, iMessage, Teams, files. Local, free.
Search, read, and write your Apple Notes from ChatGPT/Claude via a local Mac agent + MCP relay.
Connect Claude to Fathom meeting recordings, transcripts, and summaries
Persistent context for Claude. Your AI always knows your projects and next actions across sessions.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/docdyhr/macwhisper-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
