What’s new

PageToVid’s MCP server grew from 38 to 74 tools in September 2026. This page lists what was added and which tool does it. Every tool is documented in the tool reference, which is generated from the server itself.

  1. Know the price before you spend
  2. Choose the model, or let PageToVid choose
  3. Films that fix themselves
  4. Sound and languages
  5. One storyboard, every format
  6. Reuse instead of paying twice
  7. Send your own files
  8. Brand and defaults
  9. Frames and timeline
  10. Teams

Know the price before you spend

  • list_models lists every clip and still model with what it can do and its price in credits.
  • quote_cost prices one generation, or the next re-render of a video, and spends nothing.
  • dry_run: true on create_video, create_animation, rerender_video and generate_asset returns the quote instead of starting anything.
  • max_credits on the same tools refuses with budget_exceeded before anything is charged.

See Credits & pricing.

Choose the model, or let PageToVid choose

  • Name a model per clip (model), or use model: "auto" with optimize: quality, balanced, speed or cost. The reason for each choice comes back in models_used.
  • Clip controls, sent only where the model accepts them: duration_s, resolution, aspect_ratio (including 21:9), first_frame and last_frame, seed, negative_prompt, camera and shot.
  • best_of: 2–4 generates several candidates, grades each one, and keeps the best. Every candidate is charged.

Films that fix themselves

  • A number is never split across two captions, long silences are trimmed, and a caption that would leave the frame is recut.
  • quality_gate on create_video and rerender_video can hold a film under a score. release_video publishes a held film.
  • ai_mode on create_video: none, assist, or rich, where the planner may add AI stills and clips within ai_budget.
  • Each voice at its own pace. The twelve voices do not read at the same speed. PageToVid learns each voice’s pace from finished films and uses it for the word budget and the length checks, so the film you ask for is closer to the length you asked for.

Sound and languages

  • translate_video makes the same film in up to six languages. Quotes stay word for word, numbers are formatted for each language, and filmed scenes are re-recorded on the site’s own localised page where it has one.
  • Per-scene voice, speed and emotion: the set_scene_voice operation of update_storyboard.
  • pronunciations: a lexicon for the voice only. Captions keep the written form.
  • [pause 500ms] markers in narration become real silence.
  • list_voices includes a real sample of every voice in every language.
  • Arabic (captions right to left), Hindi, Korean, Polish, Turkish and Swedish narration.
  • upload_asset brings your own logo, footage or music; use a track with music: "asset:<id>".

One storyboard, every format

  • outputs: 16:9, 9:16, 1:1 and 4:5 from one storyboard, each shape as a sibling video.
  • draft: true: a free, low-resolution preview marked DRAFT, rate-limited.
  • callback_url, set_webhook and rotate_webhook_secret: a signed event when a render ends.

Reuse instead of paying twice

  • Objects (create_object, import_objects, list_objects, update_object, delete_object): a product, logo or place, reused in every generation.
  • Scene library (save_scene, list_scenes, copy_scene, and the storyboard operation add_scene_from_library): a finished scene with its footage, reused at no AI cost.
  • Characters gained update_character and delete_character.
  • Undo: list_revisions and restore_revision put a storyboard back as it was before any edit.

Send your own files

  • create_upload gives your assistant a one-hour address. It sends a file you gave it with one command (curl -sS -T <file> "<upload_url>"): a picture, a clip, music, a font or a slide deck, up to 30 MB per send. Within that, an image is at most 20 MB, a font 2 MB, and a clip 60 seconds. A deck is read up to its first 40 slides.
  • upload_asset also takes small files inline (data_base64), and reads the kind from the bytes.
  • Slide decks. Send a PDF or a PowerPoint file as kind: "deck" and every slide comes back with its text, its speaker notes and its pictures, which are added to your bank. That is enough to write the storyboard: one slide, one scene. A PDF gives an image of each slide; for a PowerPoint file, the pictures placed on the slides.
  • In the app too. Drop a file on the Media bank page. “Copy for your AI assistant” copies its ids into a sentence to paste into any chat.

Brand and defaults

  • set_brand_kit and get_brand_kit hold a site’s logo, colours, font (including an uploaded one), watermark, and default voice, music and look.
  • set_account_defaults sets your own defaults.
  • Every option resolves from account to brand kit to format to video to scene, and get_storyboard shows where each value came from.
  • list_options returns every choice in one call.

Frames and timeline

  • get_timeline gives the exact start and end of every scene, the words, and the music.
  • get_frames returns stills of the finished film.

Teams

  • Members, roles and site scopes: invite_member, list_members, update_member, remove_member.
  • One wallet per workspace, with a monthly cap per member.
  • Review links for people without an account (create_review_link).
  • Approvals: request_approval, approve, request_changes.
  • Comments and an audit log: list_comments, add_comment, get_audit_log.
  • Seats per plan: Pro 3, Scale 5, Business 15, Agency unlimited.
  • On the website too. A member picks the workspace with the switcher at the top of the dashboard. The films, sites and buttons they see follow their role and their sites, and what they spend counts against their monthly cap. Deleting a film is for the owner and admins.