{"authentication":{"bearer":{"description":"Alternative: personal API key (Settings → API or POST /api/me/api-key). Both OAuth and API key need the Starter plan or above.","get_token":"https://cliponit.app/billing","header":"Authorization","scheme":"Bearer","token_prefix":"ck_"},"description":"OAuth 2.1 (dynamic client registration or client ID metadata document + PKCE S256): Claude.ai / Claude Desktop / Claude Code / ChatGPT connect in one click from the server URL. Starter plan and above.","methods":["oauth2","bearer"],"oauth2":{"authorization_endpoint":"https://cliponit.app/oauth/authorize","authorization_server_metadata":"https://cliponit.app/.well-known/oauth-authorization-server","client_id_metadata_document":true,"dynamic_client_registration":true,"flow":"authorization_code","pkce":"S256","protected_resource_metadata":"https://cliponit.app/.well-known/oauth-protected-resource","registration_endpoint":"https://cliponit.app/oauth/register","scopes":["clips"],"token_endpoint":"https://cliponit.app/oauth/token"},"required":true,"type":"oauth2"},"capabilities":{"prompts":false,"resources":false,"roots":false,"sampling":false,"tools":true},"description":"Two products in one server: (1) turn long videos into short, subtitled, reframed clips: import (file or URL), detect viral moments, auto-reframe on faces, cut silences, subtitles, B-roll, music, emojis, export 9:16 or 16:9 MP4; (2) compose vertical AI explainer videos from a topic, a text or a voice recording (explainer_* tools, explainer_produce end to end).","documentation":"https://cliponit.app/docs","endpoints":{"streamable_http":"https://cliponit.app/mcp"},"llms_txt":"https://cliponit.app/llms-api.txt","mcp_version":"1.0","name":"io.github.MattJeff/cliponit","openapi":"https://cliponit.app/openapi.json","pricing":"https://cliponit.app/billing","remotes":[{"type":"streamable-http","url":"https://cliponit.app/mcp"}],"serverInfo":{"name":"ClipOnIt","version":"0.1.0"},"server_name":"ClipOnIt","server_version":"0.1.0","tools":[{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"List videos"},"description":"List the user's videos, newest first: id, title, status (processing|ready|failed), state (queued|running|ready|failed), queue_position, eta_s, progress (downloading|probing|transcribing|detecting while processing), duration_s (seconds), clips_count, is_demo (1 = sample video, not exportable on free). Start here to find a video_id.","inputSchema":{"additionalProperties":false,"properties":{},"required":[],"type":"object"},"name":"list_videos","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"List videos"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Upload video file"},"description":"Upload a local video file from the server's filesystem (mp4/mov/webm; path under $HOME, $TMPDIR, /tmp or data/) and start processing (transcription + clip detection). Same quota rules as the UI. Returns the video (state queued|running); then call wait_video.","inputSchema":{"additionalProperties":false,"properties":{"brief":{"description":"Free text (≤ 1000 chars): what you want, e.g. 'clips that lead to the Mansion Club, no car-only clips'.","type":"string"},"goal":{"description":"What the clips are for: views (reach), clients (sell the offer: proof, method, CTA to the offer), comments (debate), followers (story, series). Changes which moments are picked, how they are scored (goal_fit sub-score), the titles, and adds a caption CTA per clip (clip.cta).","enum":["views","clients","comments","followers"],"type":"string"},"path":{"description":"Absolute path to a .mp4/.mov/.webm file on the server","type":"string"},"resources":{"description":"Optional. {kind:'site', url:'https://…'}: the offer's site, read server-side to align the CTA with the real offer. {kind:'video', video_id}: another of your videos, usable as B-roll (illustrate with source_video_id).","items":{"properties":{"kind":{"enum":["site","video"],"type":"string"},"url":{"type":"string"},"video_id":{"type":"string"}},"required":["kind"],"type":"object"},"maxItems":5,"type":"array"}},"required":["path"],"type":"object"},"name":"upload_video_file","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Upload video file"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":false,"title":"Import video url"},"description":"Import a video from a public URL (YouTube, Vimeo, direct mp4…, via yt-dlp, max 1080p) and start transcription + clip detection. Counts against the monthly quota (HTTP 402 when reached). Returns the video (state queued|running, progress downloading); then call wait_video until state is ready (1-5 min for 20 min of video) or failed (error field). Example url: https://www.youtube.com/watch?v=e7Zl5VWejQc","inputSchema":{"additionalProperties":false,"properties":{"brief":{"description":"Free text (≤ 1000 chars): what you want, e.g. 'clips that lead to the Mansion Club, no car-only clips'.","type":"string"},"goal":{"description":"What the clips are for: views (reach), clients (sell the offer: proof, method, CTA to the offer), comments (debate), followers (story, series). Changes which moments are picked, how they are scored (goal_fit sub-score), the titles, and adds a caption CTA per clip (clip.cta).","enum":["views","clients","comments","followers"],"type":"string"},"resources":{"description":"Optional. {kind:'site', url:'https://…'}: the offer's site, read server-side to align the CTA with the real offer. {kind:'video', video_id}: another of your videos, usable as B-roll (illustrate with source_video_id).","items":{"properties":{"kind":{"enum":["site","video"],"type":"string"},"url":{"type":"string"},"video_id":{"type":"string"}},"required":["kind"],"type":"object"},"maxItems":5,"type":"array"},"url":{"description":"Public video URL","type":"string"}},"required":["url"],"type":"object"},"name":"import_video_url","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Import video url"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Set video objective"},"description":"Set (replace) the clip objective of a video: goal, brief, resources (absent field = cleared). It applies to the next detection: pass regenerate:true to re-detect the clips now with it (transcript reused, no new quota). Returns {goal, brief, resources, site_summary (the site as read, null until a detection ran), playbook_version}.","inputSchema":{"additionalProperties":false,"properties":{"brief":{"description":"Free text (≤ 1000 chars): what you want, e.g. 'clips that lead to the Mansion Club, no car-only clips'.","type":"string"},"goal":{"description":"What the clips are for: views (reach), clients (sell the offer: proof, method, CTA to the offer), comments (debate), followers (story, series). Changes which moments are picked, how they are scored (goal_fit sub-score), the titles, and adds a caption CTA per clip (clip.cta).","enum":["views","clients","comments","followers"],"type":"string"},"regenerate":{"default":false,"description":"Re-detect the clips now with this objective (replaces the current clips)","type":"boolean"},"resources":{"description":"Optional. {kind:'site', url:'https://…'}: the offer's site, read server-side to align the CTA with the real offer. {kind:'video', video_id}: another of your videos, usable as B-roll (illustrate with source_video_id).","items":{"properties":{"kind":{"enum":["site","video"],"type":"string"},"url":{"type":"string"},"video_id":{"type":"string"}},"required":["kind"],"type":"object"},"maxItems":5,"type":"array"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id"],"type":"object"},"name":"set_video_objective","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Set video objective"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get video"},"description":"Get a video (status, state queued|running|ready|failed, queue_position, eta_s = estimated seconds left, progress step, duration_s, error) with its detected clips once ready, sorted chronologically. Each clip has id, title, start_s/end_s (absolute seconds in the source), score (0-100 viral potential), reason (why), description and hashtags (null until describe_clip), cta (caption call to action, only when the video has a goal), thumbnail_url. Prefer wait_video over polling this.","inputSchema":{"additionalProperties":false,"properties":{"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id"],"type":"object"},"name":"get_video","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get video"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Wait video"},"description":"Block (server-side, up to timeout_s ≤ 120) until the video's state is ready or failed, then return it exactly like get_video (clips included). If still queued/running at the timeout, returns the video with timed_out:true, eta_s and queue_position: just call it again. Use this instead of polling get_video.","inputSchema":{"additionalProperties":false,"properties":{"timeout_s":{"default":120.0,"description":"Max seconds to block (server-side long-poll); the object is returned with timed_out:true if still in progress","maximum":120.0,"minimum":1,"type":"number"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id"],"type":"object"},"name":"wait_video","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Wait video"},{"annotations":{"destructiveHint":true,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Delete video"},"description":"Delete a video with its clips, edits, exports and files. Irreversible. Returns {ok:true}.","inputSchema":{"additionalProperties":false,"properties":{"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id"],"type":"object"},"name":"delete_video","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Delete video"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Find moment"},"description":"Find a moment in a ready video from a natural-language query (AI search over the transcript), e.g. 'the part about pricing'. Creates and returns a new clip {id, title, start_s, end_s} (absolute seconds) appended to the video's clips (delete_clip to remove it). Requires state ready. Free plan: 5 searches/day (HTTP 402 beyond).","inputSchema":{"additionalProperties":false,"properties":{"query":{"description":"What to look for, e.g. 'the part about pricing'","type":"string"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id","query"],"type":"object"},"name":"find_moment","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Find moment"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Create clip"},"description":"Create a clip by hand on a ready video from exact bounds (absolute seconds in the source, 3 ≤ end_s - start_s ≤ 180). Bounds within 0.6 s of a transcript sentence start/end are snapped to it, else to the nearest word boundary within 0.6 s (snap:false keeps them as given), then rounded to 0.01 s. Without title, the clip is returned at once with its first words as title and an AI title is written a few seconds later (get_video to read it). The clip is inserted at its chronological position (later clips move down one rank) and returned with its thumbnail_url; score stays null. Does not count against the monthly video quota. Tip: pick bounds from get_transcript word timings. HTTP 400 when out of range or the video is not ready, 404 when the video is not yours.","inputSchema":{"additionalProperties":false,"properties":{"end_s":{"description":"End, absolute seconds (3 to 180 s after start_s)","type":"number"},"snap":{"default":true,"description":"Snap bounds to sentence starts/ends (else word boundaries) within 0.6 s","type":"boolean"},"start_s":{"description":"Start, absolute seconds","minimum":0,"type":"number"},"title":{"description":"Optional title (max 80 characters); AI-generated when omitted","type":"string"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id","start_s","end_s"],"type":"object"},"name":"create_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Create clip"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get clip"},"description":"Get a clip with its edit state for an orientation: layout, keyframes [{t (seconds from clip start), layout, zones:[{x,y,w,h} fractions 0..1 of the source]}], cuts [{start,end} seconds], subtitles settings, broll, music, emojis.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"}},"required":["clip_id"],"type":"object"},"name":"get_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get clip"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Illustrate clip with video"},"description":"B-roll from the user's OWN footage: cut over the speaker, full frame, shots taken from OTHER moments of the same source video or from the account's B-roll library (all their videos) that show what is being said (e.g. the car when they talk about the car). First call on a video indexes its shots (one keyframe per shot described by a light vision model, ~30-60 s, ~0.1 $, then cached). Rules: never a shot already inside the clip, never a talking-head shot, 1.5-2.5 s each, at most one every 4 s, no overlap with existing B-roll; the voice keeps playing. Inserts are added to the clip's broll for that orientation (src:'video', source_s = where the shot starts in the source) and can be removed like any B-roll. Returns {broll (full list), inserts:[{start, end, source_s, shot, word, description}], shots_indexed}. Paid plans (403 on free).","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"}},"required":["clip_id"],"type":"object"},"name":"illustrate_clip_with_video","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Illustrate clip with video"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Broll search"},"description":"Search the account's B-roll library by words (description and tags, accent-insensitive, any language of the video + English): reusable shots extracted from ALL the user's videos (no talking head, bright and sharp enough, 1.5-6 s), e.g. q='Aston rouge nuit' or 'car driving highway'. Returns {items:[{id, video_id, video_title, start_s, end_s, duration_s, description, tags, score (0..1), favorite, vertical_url (720x1280 mp4), horizontal_url (1280x720 mp4), thumb_url}], total, videos, max}, best matches first. Then broll_add_to_clip.","inputSchema":{"additionalProperties":false,"properties":{"limit":{"default":20,"type":"integer"},"q":{"description":"Words to look for","type":"string"},"video_id":{"description":"Only shots from this video","type":"string"}},"required":["q"],"type":"object"},"name":"broll_search","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Broll search"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Broll list"},"description":"List the account's B-roll library (favorites first, then by quality score): same items as broll_search, plus videos:[{id, title, count, status: none|running|done|failed}] (extraction runs automatically after processing on paid plans). Filters: video_id, tag, favorite. max = library cap on the free plan (null = per-video cap only).","inputSchema":{"additionalProperties":false,"properties":{"favorite":{"type":"boolean"},"limit":{"default":100,"type":"integer"},"tag":{"type":"string"},"video_id":{"type":"string"}},"required":[],"type":"object"},"name":"broll_list","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Broll list"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Broll add to clip"},"description":"Place a shot of the B-roll library (from any of the user's videos) on a clip as a full-frame, muted cutaway on [start, end] (absolute seconds inside the clip; end omitted = shot length, 3 s max). Returns the clip's broll list for that orientation. Paid plans.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"end":{"description":"Absolute seconds, inside the clip (optional)","type":"number"},"item_id":{"description":"id from broll_search / broll_list","type":"string"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"start":{"description":"Absolute seconds, inside the clip","type":"number"}},"required":["clip_id","item_id","start"],"type":"object"},"name":"broll_add_to_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Broll add to clip"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"List video shots"},"description":"List the shots of the clip's source video usable as B-roll for this clip (indexes the video on first call, see illustrate_clip_with_video): {shots:[{idx, start_s, end_s, description, keywords, thumb_url}], total}. Talking-head shots and shots inside the clip are left out. Then add_video_broll to place one.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"}},"required":["clip_id"],"type":"object"},"name":"list_video_shots","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"List video shots"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Add video broll"},"description":"Place a moment of the clip's own source video as a full-frame B-roll: the source from source_s for (end - start) seconds, muted, covers the clip on [start, end] (absolute seconds inside the clip, 10 s max). shot = idx from list_video_shots (optional, labels the insert). Returns the clip's broll list.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"end":{"description":"Absolute seconds, inside the clip","type":"number"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"shot":{"description":"Optional shot idx from list_video_shots","type":"integer"},"source_s":{"description":"Where to take the footage in the source video (seconds)","minimum":0,"type":"number"},"start":{"description":"Absolute seconds, inside the clip","type":"number"}},"required":["clip_id","start","end","source_s"],"type":"object"},"name":"add_video_broll","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Add video broll"},{"annotations":{"destructiveHint":true,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Delete clip"},"description":"Delete a clip (its edits, exports and files; running exports are canceled). The video keeps its other clips. Returns {ok:true}.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"}},"required":["clip_id"],"type":"object"},"name":"delete_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Delete clip"},{"annotations":{"destructiveHint":true,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Set clip bounds"},"description":"Move the clip's start_s/end_s (absolute seconds in the source video, duration ≥ 3 s). Keyframes are shifted, cuts/B-roll/emojis outside the new range are dropped, the clip thumbnail/filmstrip is regenerated. Returns the clip. Tip: pick bounds from get_transcript word timings, or call snap_clip_to_words afterwards.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"end_s":{"type":"number"},"start_s":{"minimum":0,"type":"number"}},"required":["clip_id","start_s","end_s"],"type":"object"},"name":"set_clip_bounds","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Set clip bounds"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Snap clip to words"},"description":"Snap the clip's start_s to the nearest word start and end_s to the nearest word end in the transcript (segment bounds when word timings are missing), so the clip does not cut a word. No-op if the result would be shorter than 3 s. Returns the clip.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"}},"required":["clip_id"],"type":"object"},"name":"snap_clip_to_words","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Snap clip to words"},{"annotations":{"destructiveHint":true,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Update clip edit"},"description":"Patch the edit state of a clip for an orientation: only the given fields change, the rest is kept; subtitles is merged field by field (e.g. {subtitles:{style:'boxed'}} keeps color/size/…). keyframes: [{t (seconds from clip start), layout, zones:[{x,y,w,h} fractions 0..1], ease?: cut|smooth}] replaces the list; cuts: [{start,end}] in absolute source seconds replaces the list; broll/emojis replace their list, music: null removes it (Starter+). Returns the full stored edit. 400 names the invalid field.","inputSchema":{"additionalProperties":false,"properties":{"broll":{"description":"[{id, start, end, src, path, place?}] as returned by get_clip (absolute seconds). place {x,y,w,h} = fractions of the OUTPUT frame; omit it for a full-frame cutaway, set {\"x\":0,\"y\":0.5,\"w\":1,\"h\":0.5} with layout `duo` for the vertical duo (second video in the bottom half)","items":{"type":"object"},"type":"array"},"clip_id":{"description":"Clip id (UUID)","type":"string"},"cuts":{"items":{"properties":{"end":{"type":"number"},"start":{"type":"number"}},"required":["start","end"],"type":"object"},"type":"array"},"emojis":{"items":{"properties":{"emoji":{"type":"string"},"end":{"type":"number"},"size":{"enum":[1,2,3],"type":"integer"},"start":{"type":"number"},"x":{"type":"number"},"y":{"type":"number"}},"required":["start","end","emoji"],"type":"object"},"type":"array"},"keyframes":{"items":{"properties":{"ease":{"enum":["cut","smooth"],"type":"string"},"layout":{"type":"string"},"t":{"minimum":0,"type":"number"},"zones":{"items":{"properties":{"h":{"type":"number"},"w":{"type":"number"},"x":{"type":"number"},"y":{"type":"number"}},"required":["x","y","w","h"],"type":"object"},"type":"array"}},"required":["t","layout","zones"],"type":"object"},"type":"array"},"layout":{"enum":["vertical","horizontal","split","trio","spotlight","centered","duo","keynote"],"type":"string"},"music":{"properties":{"ducking":{"default":true,"type":"boolean"},"fade_s":{"default":1.5,"maximum":10,"minimum":0,"type":"number"},"music_id":{"type":"string"},"volume":{"default":0.25,"maximum":1,"minimum":0,"type":"number"}},"type":["object","null"]},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"subtitles":{"additionalProperties":false,"description":"Burned-in subtitles; every field optional, merged over the current values","properties":{"animate":{"default":false,"description":"Pop animation on the current word","type":"boolean"},"bw":{"default":false,"description":"Black & white grade of the whole image (recommended with behind)","type":"boolean"},"case":{"default":"normal","enum":["normal","upper","lower"],"type":"string"},"color":{"default":"yellow","description":"Accent colour (current word in karaoke, text otherwise)","enum":["yellow","white","green","cyan","pink","orange","rainbow"],"type":"string"},"enabled":{"default":true,"type":"boolean"},"font":{"default":"Inter","enum":["Inter","Montserrat","Bangers","Poppins","Arial"],"type":"string"},"hook":{"description":"Title card shown at the top during the first seconds (the hook). text null = clip title.","properties":{"duration_s":{"default":2.5,"maximum":10,"minimum":0.5,"type":"number"},"enabled":{"default":false,"type":"boolean"},"style":{"default":"card","description":"card = rounded box, plain = text with outline","enum":["card","plain"],"type":"string"},"text":{"maxLength":120,"type":["string","null"]}},"type":"object"},"lang":{"default":"original","description":"'original' or an ISO 639-1 code to translate (paid plans)","type":"string"},"overrides":{"description":"Replace the text of the segment covering [start,end] (absolute seconds; paid plans)","items":{"properties":{"end":{"type":"number"},"start":{"type":"number"},"text":{"type":"string"}},"required":["start","end","text"],"type":"object"},"type":"array"},"position":{"default":"bottom","enum":["bottom","middle","top"],"type":"string"},"size":{"default":2,"description":"52/64/80 px on a 1080x1920 output","enum":[1,2,3],"type":"integer"},"style":{"default":"karaoke","description":"behind = one giant serif word behind the person (subject mask, Premium/Montage pro), serif = same word on top, hollow = outlined caps, highlight = spoken word in a colour box","enum":["karaoke","boxed","outline","glow","minimal","behind","serif","hollow","highlight"],"type":"string"},"words_per_line":{"default":5,"maximum":8,"minimum":3,"type":"integer"}},"type":"object"}},"required":["clip_id"],"type":"object"},"name":"update_clip_edit","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Update clip edit"},{"annotations":{"destructiveHint":true,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Set default subtitles"},"description":"Set the account's subtitle preset, applied to every clip edit created from now on (new clips, new orientations, exports of untouched clips). Fields are merged over the built-in defaults ({} resets), existing edits are not changed (use update_clip_edit for those). Returns the stored preset. Call it once before export_all_clips.","inputSchema":{"additionalProperties":false,"properties":{"subtitles":{"additionalProperties":false,"description":"Burned-in subtitles; every field optional, merged over the current values","properties":{"animate":{"default":false,"description":"Pop animation on the current word","type":"boolean"},"bw":{"default":false,"description":"Black & white grade of the whole image (recommended with behind)","type":"boolean"},"case":{"default":"normal","enum":["normal","upper","lower"],"type":"string"},"color":{"default":"yellow","description":"Accent colour (current word in karaoke, text otherwise)","enum":["yellow","white","green","cyan","pink","orange","rainbow"],"type":"string"},"enabled":{"default":true,"type":"boolean"},"font":{"default":"Inter","enum":["Inter","Montserrat","Bangers","Poppins","Arial"],"type":"string"},"hook":{"description":"Title card shown at the top during the first seconds (the hook). text null = clip title.","properties":{"duration_s":{"default":2.5,"maximum":10,"minimum":0.5,"type":"number"},"enabled":{"default":false,"type":"boolean"},"style":{"default":"card","description":"card = rounded box, plain = text with outline","enum":["card","plain"],"type":"string"},"text":{"maxLength":120,"type":["string","null"]}},"type":"object"},"lang":{"default":"original","description":"'original' or an ISO 639-1 code to translate (paid plans)","type":"string"},"overrides":{"description":"Replace the text of the segment covering [start,end] (absolute seconds; paid plans)","items":{"properties":{"end":{"type":"number"},"start":{"type":"number"},"text":{"type":"string"}},"required":["start","end","text"],"type":"object"},"type":"array"},"position":{"default":"bottom","enum":["bottom","middle","top"],"type":"string"},"size":{"default":2,"description":"52/64/80 px on a 1080x1920 output","enum":[1,2,3],"type":"integer"},"style":{"default":"karaoke","description":"behind = one giant serif word behind the person (subject mask, Premium/Montage pro), serif = same word on top, hollow = outlined caps, highlight = spoken word in a colour box","enum":["karaoke","boxed","outline","glow","minimal","behind","serif","hollow","highlight"],"type":"string"},"words_per_line":{"default":5,"maximum":8,"minimum":3,"type":"integer"}},"type":"object"}},"required":["subtitles"],"type":"object"},"name":"set_default_subtitles","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Set default subtitles"},{"annotations":{"destructiveHint":true,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Auto reframe"},"description":"Detect faces/speakers, camera cuts, the audio speech level and presentation slides in the clip and write layout + keyframes for the orientation. mode=auto (default): keynote layout if a slide + a speaker are detected, otherwise speaker_tracking; mode=speaker_tracking (podcasts): vertical on the person who is TALKING (audio level x mouth movement at 10 fps), clean cut at each turn (held >= 1.4 s), split top/bottom during rapid back-and-forth; mode=keynote_split: slide on top (16:9) + speaker below (layout `keynote`); mode=keynote_speaker: speaker only, slide cropped out; mode=follow (legacy): vertical following the active speaker, split while the clip's stable duo is on screen; mode=mix (auto layout, vertical only): layout picked shot by shot — 1 face vertical, 2 faces split (or vertical on someone talking alone >= 4 s), 3 faces trio, no usable face or a group -> full-frame wide shot on blurred background. Shots without faces never get a crop on the void. framing: tight (face close-up), medium (default, face + shoulders), wide (least zoom); zoom is capped to avoid grain. horizontal = full frame. Synchronous, ~5 s per 30 s of clip; call it before export for talking-head videos. Overwrites existing keyframes (including auto_zoom ones: run auto_reframe first, auto_zoom after). Returns {layout, keyframes, faces, speakers, shots, mode (applied), slides}.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"framing":{"default":"medium","description":"How close the crop is: tight close-up, medium face + shoulders, wide least zoom","enum":["tight","medium","wide"],"type":"string"},"mode":{"default":"auto","description":"auto = keynote if slides detected else speaker_tracking; speaker_tracking = frame whoever talks; keynote_split = slide on top + speaker; keynote_speaker = speaker only; follow = legacy tracking; mix = mixed auto layout (vertical/split/trio/wide shot) per shot, 9:16 only","enum":["auto","speaker_tracking","keynote_split","keynote_speaker","follow","mix"],"type":"string"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"}},"required":["clip_id"],"type":"object"},"name":"auto_reframe","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Auto reframe"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Detect silences"},"description":"Detect silences in the clip (audio energy, 150 Hz high-pass, trimmed 80 ms each side). Without threshold_db the threshold adapts to the background noise (car, street). Returns {cuts:[{start,end,kind:'silence'}], threshold_db} in absolute source seconds; does NOT modify the edit: pass them (merged with the existing cuts from get_clip) to update_clip_edit.cuts to apply. Prefer remove_filler_words, which also cuts ums and repeats.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"min_ms":{"default":400,"description":"Minimum silence length in milliseconds","type":"integer"},"threshold_db":{"description":"Fixed silence threshold in dB (omit = adaptive)","type":"number"}},"required":["clip_id"],"type":"object"},"name":"detect_silences","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Detect silences"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Remove filler words"},"description":"Find ums/uhs (filler), hesitations, repeated words, false starts, verbal tics (aggressive only) and dead air in the clip, from a verbatim word-level transcription + audio energy. Cuts never fall inside a kept word and leave a natural breath. Returns {cuts:[{start,end,kind,confidence,text}], source, removed_s, duration_s, applied} (absolute source seconds). apply:true replaces the previous automatic cuts of the orientation's edit (manual cuts are kept); each cut can then be removed individually with update_clip_edit.cuts. ~5 s per minute of clip.","inputSchema":{"additionalProperties":false,"properties":{"apply":{"default":false,"type":"boolean"},"clip_id":{"description":"Clip id (UUID)","type":"string"},"intensity":{"default":"normal","enum":["light","normal","aggressive"],"type":"string"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"}},"required":["clip_id"],"type":"object"},"name":"remove_filler_words","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Remove filler words"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Auto zoom"},"description":"Add punch-in zoom keyframes at sentence starts (single-zone layouts only) on top of the current keyframes and save them. Run it AFTER auto_reframe (which overwrites keyframes). Returns {keyframes}.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"every":{"default":2,"description":"Zoom on every Nth sentence","type":"integer"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"strength":{"default":0.15,"description":"Zoom amount as a fraction 0..1 of the frame","type":"number"}},"required":["clip_id"],"type":"object"},"name":"auto_zoom","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Auto zoom"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Describe clip"},"description":"Generate (or return cached) a ready-to-post description and 5-8 hashtags (without #) for the clip from its transcript. Returns {description, hashtags}.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"}},"required":["clip_id"],"type":"object"},"name":"describe_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Describe clip"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Export clip"},"description":"Render the clip as MP4 with its current edit for the orientation (subtitles — account preset if the clip was never edited —, cuts, reframing, B-roll, music, emojis; watermark on free). Returns the export {id, status: queued, eta_s}; then wait_export. Demo video on the free plan: HTTP 403.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"},"music":{"description":"false = « Sans musique » : no music added by ClipOnIt (the clip's own sound/voice is kept, loudness unchanged); the user then adds a trending licensed sound in TikTok / Instagram. Remembered with the clip edit.","type":"boolean"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"style":{"default":"standard","description":"standard = the edit as is. premium (paid plans) = edited for the scroll at render time, the edit is not modified: white hook card on the first 3 s quoting a sentence actually said in the clip (else the clip title), big uppercase 3-word captions with keywords in yellow, ums and silences cut, illustration shots taken from the video, light punch-ins on keywords, no caption when nobody speaks, audio at -14 LUFS.","enum":["standard","premium"],"type":"string"},"tics":{"description":"Also remove ums, hesitations, repeats and dead air just before rendering (same as remove_filler_words apply:true; manual cuts kept).","enum":["light","normal","aggressive"],"type":"string"}},"required":["clip_id"],"type":"object"},"name":"export_clip","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Export clip"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage templates"},"description":"Montage (paid plans): list ClipOnIt's house viral formats (tier list, two-column sort, paid vs free, rated out of 10, wrong vs right split screen, before/after, symptom → cause, easy to hard, 3 mistakes, A vs B, one image per idea…). Each has an id, FR/EN name and description, and its editing recipe (JSON, schema at GET /api/montage/schema). No AI cost.","inputSchema":{"additionalProperties":false,"properties":{},"required":[],"type":"object"},"name":"montage_templates","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage templates"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage upload rush"},"description":"Montage: upload one rush (video ≤ 2 min or image, local path under $HOME, $TMPDIR, /tmp or data/) to the account's private media. Returns {id, …}: pass ids as media_ids to montage_create. Rushes stay private to the account.","inputSchema":{"additionalProperties":false,"properties":{"path":{"type":"string"}},"required":["path"],"type":"object"},"name":"montage_upload_rush","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage upload rush"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage search references"},"description":"Montage: find 6-12 real public short videos (YouTube Shorts) for a topic/format (« tier list nutrition »). Returns titles, thumbnails and links to the originals (never re-hosted). Pass one url as reference_url to montage_create: only its structure is analysed, the file is deleted. No AI cost.","inputSchema":{"additionalProperties":false,"properties":{"query":{"maxLength":80,"minLength":2,"type":"string"}},"required":["query"],"type":"object"},"name":"montage_search_references","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage search references"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage shared recipes"},"description":"Montage: editing recipes already analysed from references (shared library, structure only, nothing from any account). Reuse one as `recipe` in montage_create: 0 AI cost.","inputSchema":{"additionalProperties":false,"properties":{"family":{"description":"tiers | columns | pairs | score | versus | steps | list | demo | free","type":"string"}},"required":[],"type":"object"},"name":"montage_shared_recipes","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage shared recipes"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage create"},"description":"Montage (paid plans, counts 1 video of the monthly quota): edit the user's rushes into a vertical 9:16 video following ONE structure — template_id (montage_templates), reference_url (TikTok / Instagram Reels / Facebook / YouTube Shorts link: cuts, layout, texts, captions, tempo are measured, then the file is deleted; already-analysed references are reused for free), or recipe (JSON). intent = the user's sentence (« a demo of my app with its logo and music »). Word-by-word captions, the account logo (Settings → Logo) and our own royalty-free music are added by the export. rights must be true: only rushes the user holds the rights to. analyze_only + reference_url = just return the recipe. Then montage_wait: {status, url (mp4), edit_url (editor), report {cost_usd, warnings, shots, duration_s}}.","inputSchema":{"additionalProperties":false,"properties":{"analyze_only":{"type":"boolean"},"intent":{"maxLength":1000,"type":"string"},"media_ids":{"items":{"type":"string"},"maxItems":12,"type":"array"},"music":{"description":"none = « Sans musique »: the user adds a trending licensed sound in TikTok / Instagram; the rushes' sound is kept","enum":["auto","pulse","chill","drive","none"],"type":"string"},"recipe":{"type":"object"},"reference_url":{"type":"string"},"rights":{"type":"boolean"},"template_id":{"type":"string"}},"required":[],"type":"object"},"name":"montage_create","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage create"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage get"},"description":"Montage: current state of a montage (status queued|running|rendering|done|failed, stage, recipe, plan, report, url when done, edit_url).","inputSchema":{"additionalProperties":false,"properties":{"montage_id":{"type":"string"}},"required":["montage_id"],"type":"object"},"name":"montage_get","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage get"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage wait"},"description":"Montage: wait until the montage is done or failed (polls), then return it.","inputSchema":{"additionalProperties":false,"properties":{"montage_id":{"type":"string"},"timeout_s":{"description":"120 s max without a progress stream; call again if still running","maximum":1800,"minimum":5,"type":"integer"}},"required":["montage_id"],"type":"object"},"name":"montage_wait","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage wait"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Montage pro"},"description":"Montage pro (paid plans): a short edited like the hand-made ads — Premium render plus, by recipe: cold open on the strongest verified line said in the clip, TikTok loop with rewind (last frame meets the first; the loop already opens on the payoff, so no separate cold open), a musical beat (a 1.5-3.5 s moment WITHOUT speech from the source, chosen by loudness/motion, real sound + music), an interlude shot from the video's shot index relevant to what is said, the clip title in white boxes, split screen (speaker on top, the video's own shots below, captions on the seam), animated data cards taken from what is said (counter, table, transfer form, quote; every number checked against the transcript), ambient shots under the voice, bold uppercase captions with the keyword in orange, orange progress bar, rhythm (ums/silences cut light|normal|aggressive), hook check (score 0-3 in export.montage_report). Recipes: entrepreneur = business (cold open, split screen, data cards), podcasteur = story (title, ambient shots, musical beat, loop), formateur, ecole, marque. Natural-language requests map to arguments: « façon pub » = recipe marque, « podcast / story » = podcasteur, « formation » = formateur, « cours / école » = ecole, otherwise entrepreneur; « avec une boucle » = loop:true; « sans musique » = music:false (no music at all; the musical beat keeps its real sound). Or pass the user's words in `brief` with recipe auto. clip_id = one short (returns the export); video_id = the n best moments of the video (returns {exports}). Then wait_export; the export's montage_report says what was applied or skipped. No extra AI cost beyond the Premium render.","inputSchema":{"additionalProperties":false,"properties":{"brief":{"description":"With recipe auto: the user's request in their words","maxLength":300,"type":"string"},"clip_id":{"description":"Clip id (UUID)","type":"string"},"cold_open":{"type":"boolean"},"intermede":{"type":"boolean"},"loop":{"type":"boolean"},"moment":{"type":"boolean"},"music":{"description":"false = « Sans musique » : no music added by ClipOnIt (the clip's own sound/voice is kept, loudness unchanged); the user then adds a trending licensed sound in TikTok / Instagram. Remembered with the clip edit.","type":"boolean"},"n":{"default":3,"description":"With video_id: how many shorts (best scores first)","maximum":20,"minimum":1,"type":"integer"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"recipe":{"default":"auto","enum":["auto","entrepreneur","podcasteur","formateur","ecole","marque"],"type":"string"},"title":{"type":"boolean"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":[],"type":"object"},"name":"montage_pro","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Montage pro"},{"annotations":{"destructiveHint":false,"idempotentHint":false,"openWorldHint":false,"readOnlyHint":false,"title":"Export all clips"},"description":"Render every clip of a video for the orientation (one export per clip, same rules as export_clip; the subtitle preset applies to untouched clips). Returns {exports:[{clip_id, export_id}]}; then wait_export on each id (exports render one at a time).","inputSchema":{"additionalProperties":false,"properties":{"music":{"description":"false = « Sans musique » : no music added by ClipOnIt (the clip's own sound/voice is kept, loudness unchanged); the user then adds a trending licensed sound in TikTok / Instagram. Remembered with the clip edit.","type":"boolean"},"orientation":{"default":"vertical","description":"Target orientation of the edit/export: vertical (9:16, 1080x1920), horizontal (16:9), square (1:1, 1080x1080), portrait45 (4:5, 1080x1350). Each orientation has its own independent edit.","enum":["vertical","horizontal","square","portrait45"],"type":"string"},"style":{"default":"standard","description":"standard = the edit as is. premium (paid plans) = edited for the scroll at render time, the edit is not modified: white hook card on the first 3 s quoting a sentence actually said in the clip (else the clip title), big uppercase 3-word captions with keywords in yellow, ums and silences cut, illustration shots taken from the video, light punch-ins on keywords, no caption when nobody speaks, audio at -14 LUFS.","enum":["standard","premium"],"type":"string"},"tics":{"description":"Also remove ums, hesitations, repeats and dead air just before rendering (same as remove_filler_words apply:true; manual cuts kept).","enum":["light","normal","aggressive"],"type":"string"},"video_id":{"description":"Video id (UUID)","type":"string"}},"required":["video_id"],"type":"object"},"name":"export_all_clips","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Export all clips"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get export"},"description":"Get an export's status (queued|running|done|failed|canceled), progress (0-100 %), eta_s (estimated seconds left while queued/running), error, and its url (/files/exports/.mp4, relative to the server) when done. Prefer wait_export.","inputSchema":{"additionalProperties":false,"properties":{"export_id":{"description":"Export id returned by export_clip / export_all_clips","type":"string"}},"required":["export_id"],"type":"object"},"name":"get_export","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get export"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Wait export"},"description":"Block (server-side, up to timeout_s ≤ 120) until the export is done, failed or canceled, then return it like get_export (url when done). If still queued/running at the timeout, returns it with timed_out:true, progress and eta_s: call again.","inputSchema":{"additionalProperties":false,"properties":{"export_id":{"description":"Export id returned by export_clip / export_all_clips","type":"string"},"timeout_s":{"default":120.0,"description":"Max seconds to block (server-side long-poll); the object is returned with timed_out:true if still in progress","maximum":120.0,"minimum":1,"type":"number"}},"required":["export_id"],"type":"object"},"name":"wait_export","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Wait export"},{"annotations":{"destructiveHint":true,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":false,"title":"Cancel export"},"description":"Cancel a queued or running export (the ffmpeg render is killed). Returns the export with status canceled; 400 if it already finished.","inputSchema":{"additionalProperties":false,"properties":{"export_id":{"description":"Export id returned by export_clip / export_all_clips","type":"string"}},"required":["export_id"],"type":"object"},"name":"cancel_export","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Cancel export"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get export frame"},"description":"Grab one frame of a finished export at t_s seconds (from the start of the rendered clip) as a JPEG (longest side 360 px) to check framing and subtitles. Returns {url, bytes} plus the image inline (base64) when under 200 KB.","inputSchema":{"additionalProperties":false,"properties":{"export_id":{"description":"Export id returned by export_clip / export_all_clips","type":"string"},"t_s":{"default":0,"description":"Seconds from the start of the export","minimum":0,"type":"number"}},"required":["export_id"],"type":"object"},"name":"get_export_frame","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get export frame"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get transcript"},"description":"Get the clip's transcript segments [{start, end, text, words:[{start,end,word}]|null}] in absolute source seconds.","inputSchema":{"additionalProperties":false,"properties":{"clip_id":{"description":"Clip id (UUID)","type":"string"}},"required":["clip_id"],"type":"object"},"name":"get_transcript","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get transcript"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Get quota"},"description":"Get the user's plan and quota: {plan, videos_used, videos_limit, max_bytes, max_video_s (seconds|null), retention_days (days|null = never deleted), searches_today, searches_limit|null, watermark, logo, premiere_xml, edit_captions, translate, broll_music, api_mcp, explainers_used, explainers_limit (explainer renders this month), assist_today/assist_limit, critique_today/critique_limit, forge_today/forge_limit (daily AI requests), audio_min_used, audio_min_limit (voice-over minutes this month), source_min_used, source_min_limit (minutes of source video analyzed by the clipper this month; a video longer than what is left is refused)}. Call it first, before anything that counts against a quota (import_video_url, upload_video_file, explainer_render, explainer_produce, AI tools).","inputSchema":{"additionalProperties":false,"properties":{},"required":[],"type":"object"},"name":"get_quota","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Get quota"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Explainer styles"},"description":"List the explainer video styles: [{id, name, description, family (visual family), scene ('2d' = scenes shaped like pixel-news, '3d' = like archibald), tags, poster, theme, scene_fields, video_fields, camera_presets, catalogue, library, safe_area}]. safe_area = usable area in panel coordinates (pixel-news e.g. {x:[0,180], y:[-18,150]}: subtitles above, caption/credit below): keep every element's box inside it. Pick one, then explainer_library {style}.","inputSchema":{"additionalProperties":false,"properties":{},"required":[],"type":"object"},"name":"explainer_styles","securitySchemes":[{"scopes":["clips"],"type":"oauth2"}],"title":"Explainer styles"},{"annotations":{"destructiveHint":false,"idempotentHint":true,"openWorldHint":false,"readOnlyHint":true,"title":"Explainer library"},"description":"Read one style's explainer catalogue for an AI: common fields, animations, camera presets, transitions, then an INDEX of element names by family (~6k tokens). The library holds ~10,000 elements filed in a taxonomy; this index only lists the general-purpose set, reach the rest with explainer_suggest_elements, explainer_browse or explainer_search_elements. Recommended order: explainer_library ONCE per conversation and style -> explainer_suggest_elements for each scene (or explainer_browse) -> explainer_describe_elements for the names you keep (exact parameters and recolorable parts) -> write the script referencing elements by name (never draw or describe pixels) -> explainer_check -> explainer_render. Animated previews: /explainer-assets/