{"id":"transcription-mcp-server","name":"transcription-mcp-server","af_score":41.2,"security_score":40.8,"reliability_score":22.5,"what_it_does":"An MCP server package intended to expose transcription functionality to agent runtimes via the Model Context Protocol, allowing agents to submit audio (or references) for speech-to-text and receive transcription results.","best_when":"You want a standardized MCP tool surface for transcription inside an AI agent system, and you can provide/manage audio inputs and credentials as required by the underlying implementation.","avoid_when":"You need a fully self-contained transcription API/SDK with documented REST/OpenAPI contracts instead of MCP.","last_evaluated":"2026-04-04T21:41:18.322383+00:00","has_mcp":true,"has_api":false,"auth_methods":[],"has_free_tier":false,"known_gotchas":["Transcription requests may be non-idempotent depending on how retries are handled and whether jobs are re-submitted","Large audio inputs can hit timeouts/size limits; agent code should anticipate upload/processing constraints","If auth/credentials are environment-based, agents may fail silently without clear configuration error messages","If the MCP server wraps a third-party STT engine, rate limits and cost controls may be governed by that upstream system"],"error_quality":0.0}