Related Servers

Most tools are clearly distinct (project_, track_, label_, effect_ prefixes group by domain). However, there is notable redundancy and overlap: label_delete_audio_at vs label_cut_regions vs label_delete_regions vs label_split_regions vs label_silence_regions all operate on labeled regions with similar purposes. Similarly, cursor_to_track_end and select_cursor_to_track_end, and multiple label export functions (label_export, label_export_audio_segments, project_export_labels) create ambiguity about which to select.

The naming follows a consistent verb_noun or category_verb pattern (project_, track_, label_, effect_, edit_, select_, cursor_, transport_, generate_, auto_). There are a few inconsistencies like 'normalize' and 'compressor' and 'limiter' and 'click_removal' that lack the effect_ prefix used by sibling tools, and 'get_noise_profile'/'noise_reduction' deviate from the pattern, but overall the convention is strong and predictable.

144 tools is an extreme number for an audio editor MCP server. While Audacity is feature-rich, this surface is far beyond what agents can reasonably navigate. Many tools could be consolidated (e.g., the seven separate auto_cleanup_* pipelines, the large effect_* family, and the transcription set). This count will cause selection difficulty and cognitive overload for any agent.

The surface is quite comprehensive, covering project lifecycle (new/open/save/close/import/export), track management, selection, transport, editing, effects, generation, analysis, labeling, noise reduction, pipelines, and transcription. Minor gaps exist (no explicit record control for a specific track, no envelope editing beyond metadata, no undo/redo tools), but overall the domain is well covered for a full audio editing workflow.