Loading ToolBench
MCP server for Monet video editor providing tools to query and manipulate video projects, assets, timeline clips, and sequences
Monet provides 29 tools with mostly complete schemas and descriptions, but exhibits significant gaps in parameter documentation and error handling guidance. 18 tools (62%) have descriptions under 100 chars, which is below the 194-char baseline for A+ tools. Many parameters lack descriptions entirely (e.g., video_editor_list_assets, video_editor_get_state have empty input schemas with no guidance). Tool names follow verb_noun convention well (list_, get_, add_, search_), but several tools conflate multiple concerns (e.g., video_editor_extract_frames and video_editor_create_contact_sheet both operate on assets but lack clear differentiation). Output schemas are not documented, LLMs cannot predict what fields to expect from responses. Error handling is absent; no recovery guidance or actionable error messages are visible in the code.
Create a simple deterministic rough cut sequence from selected assets.
Return the current project graph including assets, sequences, tasks, and tool history.
List the searchable semantic segments for a specific asset.
List all imported assets and their stored metadata, transcripts, and durations.
Search imported assets using filename terms, stored tags, summaries, transcript text, and embeddings when available.
Search time-based segments across the project and return matching asset ranges.
Search within actual transcript segments. Returns an explicit unavailable state when no transcripts exist yet.
Output schemas not documented. LLMs cannot predict response structure (fields, types, nested objects). Tools like get_project_snapshot, list_project_assets return complex objects with no schema guidance.
Parameter descriptions missing or trivial. video_editor_get_state and video_editor_list_assets have empty input schemas with no guidance on what they return or when to use them.
No error handling or recovery guidance. Tools lack actionable error messages. E.g., if assetId is invalid in video_editor_get_asset_segments, no guidance on how to find valid IDs or retry.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2026-07-28+ | v2 |
Make a sequence active
Add a clip to the timeline
Add a marker to the timeline
Add a video, audio, or caption track to the active sequence
Create a contact sheet for a video asset via the local API bridge
Extract representative frames for a video asset via the local API bridge
List searchable semantic segments for a specific asset
Get current editor playhead and selection state
List all imported video assets in the project
List all clips on the timeline for a sequence
List markers for a sequence
List timeline tracks for a sequence
Move a clip to a new start time
Remove a clip from the timeline
Remove a clip and close the resulting gap across the active sequence
Insert a gap and push later clips and markers to the right
Search time-based semantic segments across the project
Set the selected asset
Set the selected clip
Set the playhead time in seconds
Split a clip at a specific time
Trim clip in point, duration, or start time
Destructive operations lack confirmation or dry-run support. video_editor_remove_clip and video_editor_ripple_delete_clip are irreversible but have no confirmation step or undo guidance.
Tool composition issues. video_editor_extract_frames and video_editor_create_contact_sheet both operate on assets but lack clear differentiation. LLMs may conflate them.