velorn (VelornLabs/velorn) is an MCP server listed on the M8ven Trust Index. It scores 54 out of 100, grade D. It declares 130 tools. No publisher has claimed this listing.
AI-native video editing built around real creative timelines, generative workflows, and local agent control.
Caution. Specific findings reduced this grade. They are listed on the page. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.
How we verified
Verified is a snapshot. Live keeps it current, and builds your track record.
⚡ Connect GitHub → continuous verification on every pushwhy connect →Who stands behind it
VelornLabs
Source: github_repo_search
Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →
Install from
The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.
These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.
get_projectReturn a concise summary of the open Velorn project, current timeline, asset counts, and MCP snapshot freshness.
get_timelineReturn the current timeline with tracks and clips. Clip timing is in seconds. Use assetId values with get_assets.
get_assetsReturn media assets from the open project. Heavy preview URLs and blobs are never exposed.
search_stock_mediaSearch Pexels photos or videos using the API key saved in Velorn Settings. Read-only. Returns result IDs, thumbnails, creator/source metadata, and opens the same results in the visible Stock tab by default. Use import_stock_media with previewOnly before downloading selected results.
import_stock_mediaPreview or bulk-import Pexels search results into the active Velorn project. Re-runs the Pexels search, validates optional result IDs, skips media already imported from Pexels by default, saves provenance, and can organize results into a Stock/Pexels/query folder. Defaults to previewOnly; applying d…
get_ai_review_passesReturn practical AI review recipes for Velorn MCP clients: timeline health, visible-shot review, hero presence checks, marker cleanup, disabled clip labeling, clip enable/disable previews, delivery checks, and current-frame questions.
get_mcp_recipesReturn practical MCP agent recipes and recommended workflows for Velorn. This is a clearer alias for get_ai_review_passes and should be used when a user asks what the MCP can do.
find_timeline_itemsSearch the active timeline and project assets for clips, tracks, markers, transitions, or assets by name, type, time range, track, color, enabled state, transforms, keyframes, or effects. Read-only.
check_media_healthCheck active project media health: missing files, zero-byte files, assets without local paths, clips referencing missing asset IDs, and unused active-timeline assets. Read-only.
inspect_export_fileInspect an exported media/XML file on disk. If ffprobe is installed, returns codec, duration, resolution, FPS, audio, and expected-target warnings. If path is omitted, tries the latest file in the project renders folder.
diagnose_comfyui_connectionDiagnose the local ComfyUI connection used by Velorn. Checks the configured localhost port, ComfyUI API endpoints, launcher state, port owner, likely install mode such as portable/Desktop/Docker/manual, and returns support-friendly next steps.
set_comfyui_connectionSet the local ComfyUI connection port used by Velorn. Supports previewOnly so assistants can propose the change before applying it. This changes Velorn settings only; it does not restart ComfyUI or edit launcher scripts.
repair_comfyui_connectionDiagnose the configured ComfyUI port, probe likely local ports, and propose or apply a settings fix when ComfyUI is reachable on a different localhost port. Use previewOnly first unless the user already approved the change.
guide_comfyui_setupBeginner-friendly Velorn-to-ComfyUI setup wizard. Diagnoses the current connection, probes common local ComfyUI ports, explains what to do for Portable/Desktop/Docker/manual installs, and can preview or apply the safe port-setting fix when ComfyUI is found on a different port.
control_comfyui_launcherPreview or apply ComfyUI launcher actions through Velorn: start, stop, or restart. Defaults to previewOnly for safety. Stop/restart can interrupt running generations and only work for ComfyUI processes owned by Velorn.
get_comfyui_launcher_logsReturn recent ComfyUI launcher log lines from Velorn, with a lightweight summary of common support issues like port conflicts, import errors, missing models/files, and CUDA memory errors.
validate_comfyui_nodesCheck whether specific ComfyUI node class names are available from /object_info. Useful when a workflow fails because custom nodes are missing or not loading.
list_velorn_workflowsList bundled Velorn workflows and graphs saved under Generate > My Workflows. Saved graphs include MCP readiness, detected Velorn markers, output type, and any setup blockers. Ready My Workflows entries can be passed to queue_prompt_generation_batch by their my-workflow: id.
inspect_velorn_workflowInspect a bundled, My Workflows, or explicit Velorn workflow JSON. Returns MCP readiness and marker requirements for saved graphs, extracts ComfyUI node classes, validates them against local /object_info, and provides setup hints.
check_export_readinessCheck whether the current timeline is ready for a standard delivery export, especially MP4 H.264 HD. Reports blockers, warnings, target codec/resolution/range, and suggested next actions.
inspect_clipInspect one timeline clip with its track, source asset, timing, transform, label, and a representative still image when available.
inspect_timeline_frameCapture the composed timeline preview frame at the playhead, a time in seconds, or a frame number. Returns visible clip context plus MCP image content when available. The result includes renderer: "live" means the frame came from the real playback pipeline (track mattes, corner pin, speed ramps, GLS…
prepare_generation_from_timeline_contextPrepare the Generate tab from the current timeline context. Uses the selected visible video/image clip when possible, otherwise the playhead frame. Defaults to previewOnly and never queues generation; applying captures the frame, opens Generate, and prefills workflow/prompt/settings for user review.
queue_prepared_generationQueue the generation currently staged in the Generate tab by using the same queue path as Velorn’s Queue button. Use previewOnly first. Applying can start local/credit-backed generation depending on the selected workflow, so only call with previewOnly=false after explicit user approval.
queue_timeline_generation_batchPreview or queue a batch of image-to-video generations from the selected clip or current playhead frame. Supports flexible variation counts per workflow, for example 2 WAN 2.2 variations and 7 LTX 2.3 variations. Defaults to previewOnly; applying can start local/credit-backed generation, so require …
list_comfyui_templatesSearch/list official ComfyUI workflow templates from the local cached/live catalog. Use this to find exact template names before queueing a template from MCP.
queue_timeline_template_generationPreview or queue an official ComfyUI template — or a community workflow imported with import_comfyui_workflow (pass importedWorkflowId) — using the selected/playhead timeline source clip as the primary input asset. Extra media inputs (e.g. a reference face image) go in assetFieldIds keyed by the man…
import_comfyui_workflowImport a community ComfyUI workflow (UI/graph export format — a comfy.org share download URL, a local .json file, or inline JSON) into Velorn as a runnable imported template. previewOnly (default) returns the dependency analysis — unknown node types, registry-resolvable node packs, model references …
install_workflow_setupPreview or run the missing-dependency install for a workflow: git-clones custom node packs into ComfyUI custom_nodes and downloads model files (possibly many GB) into the local ComfyUI models folder. Defaults to previewOnly — ALWAYS show the user the plan (node packs, model files, total download siz…
get_workflow_install_statusPoll a workflow dependency install started by install_workflow_setup: stage, current task, overall percent, completed/failed results, and whether a ComfyUI restart is recommended (new node packs need a restart to load).
generate_musicPreview or queue ACE-Step 1.5 turbo text-to-music generation on the connected ComfyUI (local model, no credits): style tags plus optional [verse]/[chorus] lyrics, exact duration 5-240s, BPM, musical key, time signature, up to 4 seed variations. Returns promptIds immediately (turbo model: seconds to …
get_music_generation_statusCheck ACE-Step music generation jobs queued by generate_music. Returns per-job status (queued/running/completed/failed) and, for completed jobs, the imported "Generated Music" asset ids/summaries ready for add_asset_to_timeline.
get_audio_analysisAnalyze the audio of a project asset (audio or video) or a timeline clip: beat grid + BPM estimate with confidence, raw onset times, loudness (peak dBFS, RMS, approximate integrated LUFS, downsampled loudness curve), and silence spans. Use it to cut to the beat of a song (pair with generate_music, t…
queue_h3_reference_videoPreview or queue one MiniMax H3 reference-to-video job from an exact Velorn image asset and audio asset. Designed for paid hero performance and lip-sync shots. Defaults to previewOnly; applying can spend Comfy credits, so require explicit approval first. No negative prompt is sent.
get_generation_queue_statusReturn live Velorn Generate-queue jobs, including queued/running/completed/failed state, progress, prompt IDs, and imported result asset IDs. Filter by workflowId, jobIds, or batchId. Read-only.
queue_prompt_generation_batchPreview or queue text-to-image/text-to-video jobs from written prompts. Supports bundled workflows, image-input workflows, and agent-ready my-workflow: ids returned by list_velorn_workflows. For saved workflows, provide any required input image through jobs[].assetFieldIds.image. Defaults to preview…
inspect_timeline_rangeSample a timeline range and return a labeled visual contact sheet/storyboard plus clip context for each sampled frame. Each sample reports renderer: "live" (true playback pipeline — mattes, corner pin, speed ramps, effects) or "legacy" (simplified fallback — basic transforms only). Live sampling nee…
inspect_visible_shotsFind top-visible shot changes in the timeline, sample near the start of each visible shot, and return a contact sheet plus shot metadata. Designed for fast-cut edit review.
get_generation_statusReturn active, failed, and recent generated asset status for the open project.
split_clipRazor timeline clips at a time. With clipId/clipIds, splits those clips; without, splits every clip under the time on all unlocked tracks (like the in-app split-all). Effects and keyframes stay on the left piece, matching the in-app razor. Defaults to previewOnly.
extract_rangeRemove a timeline time range across unlocked tracks: splits clips at the range edges, deletes everything inside, and (by default) ripples later clips left to close the gap. The core cut-down/remove-section operation. Markers are not shifted. Defaults to previewOnly.
set_clip_speedRetime clips: playback speed from 0.1x to 8x and/or reverse. Clip duration changes with speed (source span is preserved); the linked audio of a video clip is retimed with it. Defaults to previewOnly.
set_clip_audioSet audio clip mix properties: gain in dB and fade-in/fade-out durations in seconds (clamped to clip length). A video clipId resolves to its linked audio clip. To silence a clip entirely use set_clips_enabled. Defaults to previewOnly.
list_recent_projectsList recently opened Velorn projects (name, path, last modified, whether currently open). Works with no project open. Use with open_project to recover after an app restart or switch projects.
open_projectOpen a Velorn project by folder path or by recent-project name. Works with no project open; replaces the currently open project (work is autosaved). Defaults to previewOnly.
transcribe_captionsTranscribe spoken audio into timed caption cues with the local whisper caption engine (ComfyUI Qwen3-ASR only as a configured fallback). Starts a background job and returns a jobId immediately; poll get_caption_status for progress and the finished editable cue draft. Timeline scope transcribes the m…
get_caption_statusReturn the state of the active or most recent MCP caption job (transcribe_captions / generate_captions) plus the current editable caption cue draft.
update_caption_cuesEdit caption cues: fix text, retime, remove cues, or replace the whole list. Targets the transcription draft (the transcribe -> update -> generate flow) OR a live captions clip already placed on the timeline — clip edits are visible immediately and preserve each cue's styling. Default target: the dr…
generate_captionsGenerate captions from the cue draft. Timeline scope places a LIVE captions clip instantly (cues render every frame in preview and export — no baked overlay; the clip keeps its transform/grade/masks through a regenerate, and update_caption_cues edits it in place). Asset scope still renders a baked o…
get_music_video_sessionRead the complete agent-guided Music Video session from the Director workspace: song, lyrics/SRT, creative direction, cast, workflows, output settings, master and coverage passes, active plan, queue, and resumable conversation checkpoint. Read-only.
configure_music_videoPreview or update Music Video setup using existing project assets. Configures the song, lyrics, creative direction, workflows, resolution, and FPS in the same Director state used by the UI. Use import_asset_from_path first for a new local song. Defaults to previewOnly.
update_music_video_sessionPreview or persist the conversational Music Video checkpoint so an agent can resume a multi-turn creation session without hiding state from the user. Store the current phase, goal, summary, next question, decisions, approvals, and notes. Defaults to previewOnly.
manage_music_video_castPreview or manage the Music Video cast roster in the Director workspace. Add, update, remove, replace, or clear named performers and their existing image references. Defaults to previewOnly.
queue_music_video_character_assetPreview or queue a local AI portrait or character sheet for the Music Video cast. Portrait uses Z-Image Turbo; character_sheet uses Multiple Angles and requires an existing input image. Applying starts local GPU work. Defaults to previewOnly.
manage_music_video_passPreview or manage second-unit Music Video coverage passes. Create, update, remove, or activate alternate performance, environmental b-roll, and detail b-roll passes. The master pass can be activated with passId master. Defaults to previewOnly.
set_music_video_director_scriptPreview, validate, save, and optionally parse a complete Music Video director script into the editable Director plan. Targets master, active, or a coverage pass ID. Defaults to parsing and previewOnly.
update_music_video_shotPreview or update one shot in the active parsed Music Video plan. Supports prompt, timing, camera, shot type, artist, and per-shot reference overrides. Use get_music_video_plan for exact IDs. Defaults to previewOnly.
queue_music_video_keyframesPreview or queue Music Video keyframes for missing, all, or selected shots using native Director routing. Reference-free b-roll still follows Velorn local fallback rules. Applying may start local GPU work or spend cloud credits. Defaults to previewOnly.
queue_music_video_videosPreview or queue Music Video videos for missing, all, or selected shots using generated keyframes and native Director routing. Applying may start local GPU work or spend cloud credits. Defaults to previewOnly.
replace_music_video_keyframePreview or replace one Music Video Step 4 keyframe with an existing project image asset while preserving Director metadata. Use import_asset_from_path first for a new local file. Defaults to previewOnly.
replace_music_video_videoPreview or replace one Music Video Step 5 result with an existing project video asset while preserving Director metadata. Use import_asset_from_path first for a new local file. Defaults to previewOnly.
transcribe_music_video_audioPreview or run the Music Video Qwen ASR transcription/alignment workflow for the selected song. Existing SRT is only replaced when replaceExisting is explicitly true. Applying starts local GPU work. Defaults to previewOnly.
assemble_music_video_timelinePreview or assemble ready Music Video clips and song audio into a generated edit timeline with coverage tracks and vocal-performance sync locks. Existing assembled shots are kept. Pass a new timelineName to build a clean assembly without altering an older edit. Defaults to previewOnly.
replace_music_video_timeline_shotPreview or replace an already assembled Music Video timeline shot with its latest or a specified generated video, preserving timing, transforms, effects, and sync lock. Defaults to previewOnly.
save_projectPreview or explicitly save the current Velorn project, including Director state, assets, and the active timeline. Defaults to previewOnly.
get_music_video_statusSummarize Velorn music-video workflow assets, assembled clips, and sync-locked clips in the current project.
inspect_music_video_keyframeInspect one Music Video Step 4 shot through the open Generate workspace. Returns its prompt, current keyframe, generation history, active job, selected workflow, expected workflow routing, and whether it can be regenerated. Read-only.
regenerate_music_video_keyframePreview or queue regeneration of one Music Video Step 4 shot using the exact active keyframe settings and native Velorn routing. This includes reference-free b-roll fallback routing. Defaults to previewOnly and may start local GPU work or spend cloud credits when applied.
get_music_video_planReturn the parsed Music Video director plan from the open Generate workspace, including every sceneId and shotId, timing, shot type, keyframe and motion prompts, selected workflows, active jobs, and Step 4/5 completion state. Includes shots that have not generated assets yet. Read-only.
inspect_music_video_videoInspect one Music Video Step 5 shot through the open Generate workspace. Returns its input keyframe, motion prompt, timing, active workflow settings, active job, latest video, output history, and latest poster image when available. Read-only.
regenerate_music_video_videoPreview or queue regeneration of one Music Video Step 5 shot using its generated keyframe and the exact active video settings and native Velorn routing. Defaults to previewOnly and may start local GPU work or spend cloud credits when applied.
analyze_timelineReturn an AI-friendly read-only timeline health report with likely export risks, missing media, tiny clips/gaps, overlaps, track state, XML imports, transforms, and next actions.
analyze_music_video_workflowReturn an AI-friendly read-only health report for the music-video workflow, including generated keyframes/videos, failed or active jobs, assembled clips, sync locks, reruns/replacements, and next actions.
set_clip_label_colorSet or clear label colors for current timeline clips by explicit clipIds or a safe built-in filter. Non-destructive and undoable in Velorn.
set_clips_enabledEnable or disable current timeline clips by explicit clipIds or a safe built-in filter. Use previewOnly first for AI-assisted edit decisions. Undoable in Velorn.
add_timeline_markersAdd one or more labeled timeline markers at explicit times, frames, or the current playhead. Non-destructive and undoable in Velorn.
remove_timeline_markersRemove timeline markers by ID, all markers, marker color, label text, or time range. Undoable in Velorn; use previewOnly for safety.
set_timeline_marker_propertiesRename, recolor, or move existing timeline markers by ID, color, label text, or time range. Undoable in Velorn; use previewOnly before changing many review markers.
undoUndo the latest Velorn timeline edit or sequence/project-structure edit. Scope can be auto, timeline, or project. Low-risk and mirrors the app undo stack.
redoRedo the latest Velorn timeline edit or sequence/project-structure edit. Scope can be auto, timeline, or project. Low-risk and mirrors the app redo stack.
set_playheadMove the active timeline playhead to a seconds value, timecode, or frame. Useful before inspecting frames, staging generation, or setting ranges.
select_clipsSelect timeline clips by ID, current selection, filter, track, time, type, label color, or name search. Can optionally move the playhead to the first selected clip.
select_assetsSelect/preview project assets by ID, name, type, folder, status, or latest match. Velorn currently previews the first matched asset.
create_project_checkpointCreate an in-memory MCP safety checkpoint of the open project/timeline/assets for this app session. Use before risky AI edits; this does not create a saved project copy on disk.
restore_project_checkpointPreview or restore an in-memory MCP project checkpoint created during this Velorn session. Defaults to previewOnly because it can replace current timeline/asset state.
import_asset_from_pathPreview or import a local media path into the active Velorn project. Single files use the same copy/import path as the UI import button. Image sequences import too: point at a folder of numbered frames (or any one frame of a 3+ frame run — png/jpg/webp/tif/exr/dpx) and the run transcodes once into a…
relink_assetPreview or relink an existing Velorn asset record to a different local file path. Useful after check_media_health finds missing media. This updates Velorn project metadata only; it does not copy media. Defaults to previewOnly.
set_clip_stylePreview or batch-update simple clip styling: label color, enabled state, transform fields, crop, blur, blend mode, motion blur settings, track matte, motion path mode, corner pin, and bypass pills (per-group mask/color/effects A/B switches). Use for broad AI timeline polish passes. Defaults to previ…
set_clip_maskPreview or set clip masks (video/image clips): parametric shapes (rectangle/ellipse/rounded), bezier spline masks, or an AI mask image — the same three modes as the Inspector Mask section. Whole-mask animation goes through set_clip_keyframes (shapeMask.* properties). Defaults to previewOnly.
run_mcp_action_planPreview or run a small ordered batch of approved Velorn MCP write actions. Use this when an agent has already planned several safe steps and wants one checkpointed apply pass. Defaults to previewOnly.
set_in_out_rangeSet, preview, or clear the active timeline In/Out range. Useful before export_timeline/export_delivery_batch when the user asks for only part of the edit.
create_projectPreview or create a new Velorn project in the configured Projects folder. Applying opens the new empty project. Defaults to previewOnly so agents cannot accidentally create folders without approval.
duplicate_projectPreview or duplicate a Velorn project folder. Defaults to the current open project, copies the whole folder, remaps saved paths, creates a sibling copy, and opens the duplicate on apply.
create_timelinePreview or create a new Velorn sequence/timeline with a name, optional dimensions, fps, duration, and optional switch-to-new-sequence behavior. Defaults to previewOnly; applying is undoable in Velorn.
create_asset_folderPreview or create Velorn asset-panel folders. Accepts a single folder name or a nested path like "AI Spots / July 4th Demo"; reuses existing matching folders by default and creates only missing folders.
move_assets_to_folderPreview or move existing Velorn assets into an asset-panel folder. Can target explicit asset IDs/names or safe filters such as rootOnly + constantsOnly. Defaults to previewOnly and can create a missing target folder path before moving.
move_unused_assets_to_folderPreview or move unused project assets into an asset-panel folder without deleting files. Checks all saved project timelines plus the live active timeline, then moves only assets not referenced by any clip/effect metadata. Defaults to previewOnly.
add_trackCreate a new timeline track. Video tracks are added at the top of the stack, which is useful before adding another text/title layer. Undoable in Velorn.
update_trackPreview or update an existing timeline track: rename it, mute/unmute, solo, lock/unlock, show/hide, set mixer volume/pan, edit insert effects, change audio channels, or reorder within its video/audio group. Undoable in Velorn.
set_master_audioPreview or update the mixer master bus: program master volume and master insert effects (compressor/limiter/reverb). Both apply to preview and export. Read current master state from get_timeline (masterAudio). Track-level mixing (volume/pan/solo/inserts) lives on update_track.
30 further tools are not listed here. The complete surface is in the source.
Disclosed vulnerabilities in this server's declared npm dependencies (via OSV). Whether each is reachable depends on the installed versions.
JSONata vulnerable to Arbitrary Code Execution via crafted JSONata expressions
JSONata vulnerable to Arbitrary Code Execution via crafted JSONata expressions
JSONata: Arbitrary Code Execution via crafted JSONata expressions
jsPDF has HTML Injection in New Window paths
js-yaml: maxTotalMergeKeys does not limit CPU use for empty merge sources
ELECTRONSystemRootRUNNER_TEMPTool annotations
No tools have read-only/destructive annotations
Add readOnlyHint or destructiveHint annotations to every tool so hosts can warn users before invoking.
All four hints declared on every tool
130/130 tools missing one or more hints — get_project (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_timeline (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); get_assets (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +127 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.
For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.
Tool test coverage
Only 6/130 tools referenced in tests (5%)
Write tests that reference each tool by name so every tool has at least one test.
Production dependencies are patched
4 critical, 5 high severity in production deps — jsonata@2.1.0 (critical), jsonata@2.1.0 (critical)
Run npm audit fix, or upgrade the affected packages to a non-vulnerable version.
Dev dependencies
10 critical/high in dev-only deps (does not ship to users)
Upgrade dev dependencies when convenient.
Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.
[](https://m8ven.ai/mcp/velornlabs/velorn)?variant=verified from the URL.Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.
https://m8ven.ai/api/mcp/tool-check