Clypmint

Apply exact edits without a Clypmint model round. Read get_project_context first. Canvas ops: output_frame {frame: {ratio: 9:16, 16:9, 1:1 or 4:5, fit}} reshapes this video ("make it vertical"); fit subject (the default) follows the subject, contain keeps the whole picture with bars (only when they ask to fit or keep everything in), cover fills; optional x/y (0-1, with fit cover only) move the centre of the filled frame; add: true makes the shape an extra output and keeps this one ("a square version too"); frame: null alone (no second op) is the original shape again ("undo the crop"); layout moves/scales/rotates a returned element ID using its units; element changes supported text/color/hidden; put_overlay adds or updates timed text/images with none/fade/slide/pop animation; remove_overlay removes an added element. Use only returned editable capabilities and existing canvas image assetIds. These are the manual editor operations and preserve authored animation. Music: {op:"music",enabled:false} removes only added music; enabled:true with track (one of Clypmint's own tracks, never claimed when posted: bright-house (upbeat), lofi-desk (calm), tech-forward (tech), documentary-pulse (story), gym-phonk (energetic), acoustic-morning (warm), soft-piano (emotional), funk-pop (fun), night-suspense (mystery), rise-up (inspiring)) or with music_staging_key from import_asset (a song of their own). Optional level_db (-36 to 0 relative to source audio) and beat_sync:true for source-shot action montages; other editing clocks refuse beat sync explicitly. Scene ops: remove, remove_layer, replace_screen, set_text, swap_picture (an existing picture key), move, add, cut_words, emphasis, captions, sound, scene_sounds. remove and cut_words take footage time out of the video; to delete a graphic and keep the footage and captions under it use remove_layer {layer: graphic}, to delete a zoom or camera move remove_layer {layer: camera}, and to shorten or lengthen a graphic move its scene to fewer or more words. scene_sounds replaces the scene cue list (empty removes added cues): choose names from scenes.soundOptions, optional on from soundWords.index and gain -36 to -10 dB. It preserves speech; global sound mute still applies. Batch up to 40 ops: all apply or all refuse. Native canvas changes save a playable project; export separately for a video file. Older timeline operations may require rendering; unsupported native operations refuse without rebuilding the project. Set base_version and output_id from context. Hook openings use revise_video. Free, up to 30 a day. Preserve operation_id and exact arguments on retries. Cost: Free: no credits.

Build a reusable editing style from one reference video shared by link or file. mode clone (the default) is the deep copy: it measures the reference and writes its own captions and motion; the linked account includes style copies every month at no credit cost, and once they are used each costs 100 credits. mode match is the fast copy: it maps the reference onto Clypmint's own devices in a few minutes, is included for every account and is approximate. A copy that is still working or has stopped shows in browse_styles (with its time left, or why it stopped and how to try again). dry_run: true answers whether this one is included (and how many are left) or what it costs, and starts nothing.

Make a faceless video without filming or uploading footage. Supply exactly one of idea (Clypmint writes the narration) or script (spoken exactly as supplied). Defaults to a 30-second vertical video; seconds may be 15, 30 or 60. An exact script sets its own length at about 2.5 words a second (at most 172 words for 60 s); leave seconds out with a script and the shortest length that fits is used, and a script too long is refused at once with the word count. Uses the selected style’s pictures, designed screens, captions and sound, with the linked account’s existing credits and limits. Style: inspect browse_styles and its actual inputs. Serif Editorial Launch is format: motion with inputs.voice:none by default; Bold Navy Statement is narrated faceless with Chris as its style default. An explicit voiceover request adds narration to motion while keeping its visuals and aligning them to the recorded speech. Other faceless styles or ready saved copies remain available. A style marked format: talking is made for someone on camera. Optional: style_id (with style_version for a saved style of your own), formats (9:16, 16:9, 1:1, 4:5; first is the main one), voice (one of 21 narrators by name; omitted, preserve the style’s voice default, including none for motion; "Voice: <name>" in instructions works too), sound_effects (on unless false), music_bed (the style’s music under the voice, only when true), purpose, instructions up to 8,000 characters, asset_ids from list_assets, asset_staging_keys from import_asset (logo: true for the brand’s logo), and music_staging_key from import_asset. Cost: 3 credits to make, plus 1 for each started minute of voice-over and 1 for each started minute of finished video after the first, plus 2 for each extra format; dry_run: true answers the cost and starts nothing. For several videos at once use create_videos; for a month of posts use plan_content. Return the studio_url and call check_video with video_id for progress and the finished download. Preserve operation_id and arguments on retries; never publish unless separately requested.

Make up to 10 faceless videos in one call (more: batches of 10; a month of posts is plan_content), each from an idea or an exact script, with shared defaults (seconds, style_id, formats, voice: matilda/sarah/jessica/alice/liam/chris/brian/george/roger/laura/charlie/river/will/eric/bella/daniel/lily/bill/adam/callum/harry, sound_effects). Every item is checked before any starts; each is idempotent on its own operation id (operation_id.N), so a retry never makes doubles. Each video costs 3 credits to make, plus 1 for each started minute of voice-over and 1 for each started minute of finished video after the first, plus 2 for each extra format; dry_run: true answers the total and starts nothing. Returns the video ids; check each with check_video, then schedule_plan. Cost: About 4 credits per 30-second video (3 credits to make, plus 1 for each started minute of voice-over and 1 for each started minute of finished video after the first, plus 2 for each extra format).

Change ideas by number (n) before they are made or posted: title, hook, script, caption, hashtags, posting time, platforms, or removed: true to drop one (false puts it back). Words cannot change once an idea is being made; a scheduled post changes with update_post. Free. Cost: Free: no credits.

Edit the user's own footage into a finished video. Put what they asked for into the fields, never into prose rules: kind (montage: pictures and music lead, e.g. sport, travel, events; talking: someone speaks), length ({min_seconds, max_seconds} for "45-60 s"), style_id (read browse_styles once first: a saved style of theirs (own: true) that fits the request is the one to use), music_staging_key (a song from import_asset) or song_upload: true (they have a song file but no link: the upload link takes it too), camera_sound, formats (only when they asked for a shape; otherwise the video keeps its own and nothing is cropped). instructions holds only anything else they asked for, in their own words. Clypmint watches the footage itself: do not describe the footage, write rules or ask the user how to edit. In ChatGPT attach the video as video. Otherwise leave it out: the answer has a private upload link (15 minutes) to give the user, and editing starts when the video lands. One upload per video: to change the length, song, cut or shape of a video already made, use revise_video or apply_edits on that video (a shape is one apply_edits output_frame op), never a new upload; a separate new edit of the same footage is footage_from: that video_id. Like another video? reference_url (+ reference_mode clone or match) copies that video's style beside the upload, saves it to the user's styles, and the edit waits for it; leave style_id out. No footage (an idea or a script)? Use create_video. Also: purpose (organic, ad, youtube, sales_page), hooks 1-3 (an ad's openings), shorts 0-10 or 'all', clips (a long recording cut into clips alone), speaker_name, names, spellings, sound_effects, asset_ids, source_url (a public link where links are open). Cost: 3 credits to make, plus 1 for each started minute of footage and 1 for each started minute of finished video after the first, plus 2 for each extra hook, length or format and 2 for each clip; dry_run: true answers the cost and starts nothing. Preserve operation_id on retries.

Fix caption words or the key-word colour of a finished video, without a new edit; timing never moves. Either rewrite lines (from, to, text from get_project_context captions) or swap words everywhere with replace: [{ heard, spelled }] (brand names, people). A swapped name is also remembered for later videos. Each spoken word keeps a caption word, so a 2-word name cannot become 1 word here: use remember plus revise_video. Free; 5 a day per video. Cost: Free: no credits.

Forget one remembered preference by its id from list_memory. Later videos leave it out. Cost: Free: no credits.

Import one image or song up to 32 MiB for a video: a file attached in ChatGPT (file), or a public https link to the file (url) from Claude or any host that cannot attach files. logo: true marks the image as the brand’s logo. Returns an asset_staging_key for create_video’s asset_staging_keys (or music_staging_key for a song), usable for seven days. Repeating the same operation_id and file returns the original receipt. Does not import video footage (edit_video does) or read the host’s wider library. No link and no attachment? The account owner can add the logo once on the Brand page (no website needed): https://clypmint.com/studio/brand.

Start a content calendar for a topic: count faceless videos (1-31) over days (7-31), each with a title, hook, the exact script, caption, hashtags and a posting time. Best: write the ideas yourself in the conversation and send them as items (taken as written, at once); otherwise Clypmint writes them (this call keeps going until they are written, or tells you to call again with the same operation_id). Makes nothing and spends nothing; answers plan_id, the ideas and the quote (credits a video, room left this month). Show the user the ideas, change them with edit_content_plan, and approve_plan only after their yes. Use for "start a faceless account" or "a month of posts about X". Cost: Free to write; making each video costs about 4 credits (30 s).

Teach Clypmint one lasting preference for every later video of this brand, in plain words: caption size, pace, sound effects, colours, "own footage first", or how a name is spelled. Up to 200 characters. Free. Cost: Free: no credits.

Make an earlier version of a video live again (undo). Nothing is deleted. Free. Cost: Free: no credits.

Change a finished video against its saved base_version. A change asked for in words costs 1 credit (dry_run: true answers the price and sends nothing); exact edits with apply_edits and fix_captions are free. Name output_id for one hook/Short/format; attachment changes require one output when the video has several. Attach up to four still images with explicit intent: reference_only studies the image without placing it, use_in_video must place it. Supply each as a genuine ChatGPT file descriptor, an import_asset asset_staging_key, or an owned imported asset_id. PNG/JPEG/WebP may be placed; GIF references inspect the first frame only. Prepared owned video sources may be placed by ID; motion/audio analysis of video references needs a separate preparation path. Unsupported requests return a reason and preserve the saved project. Exact supported changes use apply_edits or fix_captions. On the legacy rendered revision path only (a video without a saved native project), replace retained main scratch narration when check_video.narration.can_change is true: pass explicit voice plus a voice-only change such as 'Change narrator to Brian. Keep the script and edit.' using the selected catalog name. This preserves script/edit/assets; send other edits separately. Uploaded speech, silent motion and mixed voice/visual instructions are refused on that replacement path. To ADD narration to an eligible silent main motion video on the legacy path, keep voice omitted and begin change with 'Narrator: liam Narration: <exact one-line voiceover script> <visual and sound-effect revision notes>'. Use a catalog narrator name; this supports one main hook, aligns rebuilt visuals to the recorded narration word clock, and refuses mismatched, clipped or missing voiceover. The saved native project path currently returns typed project_update_unsupported for voice replacement or narration addition because its separate narration preparation path is unavailable. Do not promise native narration support or retry by forcing a legacy rebuild. Preserve operation_id, base/output, roles and arguments on retries; a refreshed file download URL for the same host file may resume its upload. Use a new operation_id for a deliberate new change and check the resulting version. No movie export is started by a native project update. A change by typing runs in the background: this answers at once with a job_id, and check_video with that job_id shows the steps as they happen and then the editor's reply and what it could not do.

Keep a finished clip (or hook, or film) of a video as a reusable asset. Its asset_id goes in asset_ids of edit_video or create_video, so a clip can open another video, sit inside one, or become the whole video. Read output_id from check_video. Free. Cost: Free: no credits.

Change one of the linked account's saved styles in plain words, for example 'bigger yellow captions, no zooms' or 'open on the strongest line'. Saves it as the style's next version; every later video in that style follows it, and earlier versions are kept. Preserve operation_id on retries.

Check a video and show it. While the upload or the edit is still running, this waits for news before answering (up to 20 seconds; 50 in Claude), so call it again right away only if the user is waiting; the video card updates itself. state is waiting (give the user the upload link), uploading, editing, ready, draft (a rough cut, not the finished edit: say so and offer next_step) or failed (say problem and next_step). It also says when the length differs from the length asked. Download links are made only when asked: with_download: true (when the user wants the file).

Show ready styles/copies with content format (faceless/motion/talking), frame formats, jobs, best_for, actual inputs and saved style_version when available. Motion is faceless type/music with inputs.voice:none by default; an explicit voiceover request overrides that default while keeping the visual style and aligns visuals to the recorded speech. Narrated faceless styles use inputs.voice:narrator. Also returns all 21 narrator descriptions, accent/tags and public audition URLs; use catalog names in create_video voice. Recommend suitable returned styles, never invent private IDs. Listen/show auditions before selecting an undecided voice.

Return only the stable identity and display labels of the account represented by this request. Used for account linking. For account rights, credits and usage use get_profile.

Read a calendar from plan_content: each idea’s state (draft, making, ready, needs_connect, needs_time, scheduled, posted, failed), its video and post, and the quote. Cost: Free: no credits.

Return the Clypmint profile linked to this request, with its credits (used and included this billing month, and when they renew) and videos used, so you can say what a request will use before spending.

Read an owned video’s version, scenes, captions and editable canvas element IDs/capabilities. Optional output_id selects a clip/format; element_offset pages through the canvas inventory. Use the returned version/output and actual IDs with apply_edits for direct edits without a second Clypmint model round. Canvas image IDs name assets already in this version. Returns project data without a visual card.

Read every round of changes, direct edit and caption fix on a video, oldest first, with who asked, the status and the version it made. Cost: Free: no credits.

While a video is being made, read what the editor is doing: its stage, minutes left, the paper edit, each step (looked at footage, searched, previewed) and the names it heard. Use it to narrate progress instead of waiting silently. Cost: Free: no credits.

Fuller notes on using Clypmint, when a request is unusual: overview (which tool and fields for a request, and one-call changes), music, styles, editing (changes and the canvas), publishing (posting and monthly plans), capabilities. A normal request needs none of it: the tool descriptions are enough. Read-only, free.

List reusable media owned by the linked account, including the asset IDs accepted by video creation.

List the preferences, looks and spellings Clypmint remembers for the brand, with the ids forget takes. Cost: Free: no credits.

List the versions of a video, newest last, and which one is live. Use with restore_version to undo. Cost: Free: no credits.

List up to 20 recent videos in the linked Clypmint account.