ParrotScribe MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PSCRIBE_PATH | No | Path to the pscribe executable. Only needed if pscribe is not in your PATH (edge case). | pscribe |
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 | {} |
| prompts | {} |
| resources | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| pscribe_startA | Start real-time audio transcription (microphone and system audio). Returns status message confirming transcription started. |
| pscribe_stopA | Stop (pause) the current transcription session. Returns status message confirming transcription stopped. |
| pscribe_statusA | Get the current status of the ParrotScribe service. Returns: version, status (listening/stopped), session ID, duration, model, capture settings, output directory. |
| pscribe_tailA | Get recent transcript entries. Returns TOON format: timestamp,source,status,segment,confidence,duration,language,text
Use 'n' for last N lines, OR 'since_line' to poll from a specific line (mutually exclusive). For polling: call with since_line, note the last line number, then call again with that number + 1. |
| pscribe_sessionsA | List recent transcription sessions. Returns: session ID, date, duration, segment count for each session. |
| pscribe_newA | Force start a new transcription session. Use when you want a clean session boundary. Returns confirmation with new session ID. |
| pscribe_catA | Display complete transcript sessions with time-based filtering. Returns TOON format: timestamp,source,status,segment,confidence,duration,language,text
Use this for historical queries like "summarize yesterday's standup" or "what was discussed last week". Convert natural language time references to ISO8601 (e.g., "yesterday morning" → appropriate timestamp). |
| pscribe_grepA | Search for patterns across transcript sessions. Use this for queries like "did anyone mention deployment last week?" or "find all references to the API". Returns matching lines with session ID prefix. Use --count for summary statistics. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| summarize-session | Summarizes the current or a specific transcription session. |
| monitor-call | Sets up instructions for polling the transcript to monitor a call in real-time. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| Live Transcript | The live transcript of the active ParrotScribe session (last 100 entries in TOON format). |
TDQS
Scored across 8 tools
Each tool has a distinct purpose: start/stop control, status queries, live tail, session listing, forcing new sessions, full transcript display, and search. Even tail vs cat are differentiated by time scope (recent polling vs historical full sessions).
All tools share the pscribe_ prefix and use lowercase with underscores, but the suffix style mixes verbs (start, stop, tail, cat, grep) with nouns (status, sessions) and an odd 'new'. The prefix keeps it predictable, though a stricter verb_noun pattern would be more consistent.
8 tools is well-scoped for a transcription service covering control, live monitoring, session management, and historical queries. Each tool earns its place without redundancy.
The surface covers the full transcription lifecycle: start, stop, status, live polling, session listing, new session creation, full transcript viewing, and search. Minor gaps like explicit pause/resume or export/delete exist, but agents can work around them (start resumes, grep/cat serve export needs).