Modisign MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MODISIGN_NO_TELEMETRY | No | Turn off anonymous usage counts by setting MODISIGN_NO_TELEMETRY=1. |
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 | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_overlaysA | Find animated overlays by topic, family or niche. Returns compact preset choices and editor links. Reuse the same filters and pass the returned next_cursor as cursor for another page. |
| get_overlayB | Inspect an overlay preset's editable fields, types, allowed values and limits, looks, aspects, duration and free status before configuring it. |
| configure_overlayB | Configure an animated overlay with editable values, look, aspect, duration or background. Returns its configuration URL and free/Pro tier. Field errors list allowed inputs so you can correct and retry. Changes to free preset text generally need Pro to render. |
| render_overlayA | Start a local animated overlay render and return job_id. MP4 uses green screen by default; MOV is transparent ProRes 4444; WebM is transparent VP9. Needs the user's FFmpeg. Free configurations need no login; others need Pro. Poll job_status for the file. |
| add_overlay_to_videoA | Render an overlay transparently and composite it onto one local video (up to 60 seconds) at start_seconds. Automatically selects and reauthorizes the nearest supported aspect, returned as aspect. The start must fit the overlay within the video. Returns job_id; poll job_status. Produces H.264/AAC MP4 with audio, source aspect, long side capped at 1920. Local media never uploads. Use edit_video for longer or multi-clip edits. |
| edit_videoA | Create a local video with sequential v1 clips and nonoverlapping v2 clips above them. Every clip requires track; v2 requires start seconds, v1 forbids start. Crop uses source-edge percentages; place selects full, halves, corner picture-in-picture or a custom normalized box. Omitted crop/place mean zero crop/full. Supports trims, fill/fit, pan (0–1), zoom (1–3), photos, audio and source-second punch-ins. Up to 300 clips, maximum end 180 seconds. after_clip indexes the request array and must target v1; the incoming neighbor is the next v1, skipping v2. Transitions clamp to adjacent clip limits. Captions and pause detection use v1 only, even when muted. Removing pauses ripples both tracks and overlays; conflicting v1 transitions return an explicit error. Align sources manually; no automatic sync or drift correction. Example: put reference.mp4 on top muted, me.mp4 on the bottom, crop 5% off the sides, start together, Clean captions from my voice. Captions are made on your computer in English; optional corrections replace exact words/phrases. Specify caption style clean/boxed/minimal/karaoke/bold-keyword/word-pop/highlight-box/neon/pop-up/typewriter/outline/editorial/gradient, position upper/center/lower and size s/m/l. No overlays or unchanged free overlays with Clean/Boxed/Minimal captions export without login; other edits need Pro. 1080p default, 720p also free. Media never uploads; Studio exports never enter account history. Returns job_id; poll job_status. |
| get_transcriptA | Transcribe English speech in one local video (up to 180 seconds) using an English speech model on your computer. Optional in/out trim seconds; returned line times are source seconds. The transcript text is returned to your assistant/provider, so call only when the user wants to review or correct it. Media and transcript are never sent to Modisign. First use downloads the speech model; later calls work offline. Use captions.corrections in edit_video to fix names/phrases. |
| job_statusA | Check a local video job's state and progress (0–1). A done job includes the local artifact path, MIME type, bytes, duration and any history warning. Jobs last for this server process. |
| cancel_jobB | Cancel a queued or running local video job and clean its temporary files. A completed artifact is retained. |
| account_statusA | Check whether Modisign is signed in and has current Pro access. Anonymous free overlays work without signing in. Returns sign-in instructions when needed. |
| loginA | Start Modisign device sign-in. Show the returned verification URL and code to the user; they sign in with Google and explicitly approve. Polling continues in the background. Use account_status to check completion. This tool never opens a browser or returns credentials. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 11 tools
Each tool targets a distinct resource/action: overlay discovery (list/get), configuration, rendering, compositing, video editing, transcription, job lifecycle, and auth. The only potential overlaps—render_overlay vs add_overlay_to_video and add_overlay_to_video vs edit_video—are explicitly clarified by descriptions indicating when to use each.
8 of 11 tools follow a clear verb_noun pattern (get_overlay, cancel_job, list_overlays, configure_overlay, render_overlay, add_overlay_to_video, edit_video, get_transcript). job_status/account_status use noun_status and login is a bare verb, but these deviations are minor and readable.
11 tools is well within the ideal 3-15 range for an overlay/video-editing server. Each tool covers a necessary slice: discovery, configuration, rendering/compositing, editing, transcription, job control, and auth.
Core workflows are fully covered: overlay search/inspection/configuration/render/composite, comprehensive local video editing with captions, transcription, job status/cancel, and login/account checks. Minor gaps remain (no media metadata probe, no way to list saved overlay configurations, no logout), but agents can work around them.