Brand Kit

The Brand Kit is what makes generated output look like it came from your workspace rather than a generic template: music, opening and closing scenes, watermark, background, caption and article styling, default voice and language.

These defaults were always applied to API-generated videos - but nothing could read or change them, so an API-only integration was stuck with whatever the dashboard had been set to. Now it isn't.

Everything here needs brand_kit:read, or brand_kit:write to change.

Get the Brand Kit

GET /v1/brand-kit

curl https://api.vidocu.ai/v1/brand-kit \
  -H "Authorization: Bearer vdo_live_your_key_here"

{
  "brandKit": {
    "companyContext": "Industrial label printers for pharmaceutical packaging",
    "defaultLanguage": "en",
    "defaultVoiceId": "pNInz6obpgDQGcFmaJgB",
    "openingSceneUrl": "https://storage.vidocu.ai/...",
    "backgroundMusicUrl": "https://storage.vidocu.ai/...",
    "backgroundMusicVolume": 70,
    "watermark": { "enabled": true, "position": "bottom-right", "opacity": 70 },
    "logoUrl": "https://storage.vidocu.ai/..."
  },
  "counts": { "skills": 2, "glossary": 34, "pronunciations": 6 }
}

Skills, glossary and pronunciations are returned as counts, not contents - each has its own endpoint, and a Brand Kit with a 200-term glossary shouldn't ship it on every read.

Update the Brand Kit

PATCH /v1/brand-kit

Only the fields you send change. Send null to clear one.

curl -X PATCH https://api.vidocu.ai/v1/brand-kit \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"backgroundMusicVolume": 45, "openingSceneUrl": null}'

caution

These apply to every future export and article in the workspace, not just your next request. For a one-off change, use the per-request export overrides instead.

Pronunciations

GET /v1/brand-kit/pronunciations · PUT /v1/brand-kit/pronunciations

How specific words should be said in generated narration - product names, acronyms, anything a text-to-speech model reliably gets wrong.

curl -X PUT https://api.vidocu.ai/v1/brand-kit/pronunciations \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"pronunciations": [
    {"word": "Vidocu", "pronunciation": "vid-oh-koo"},
    {"word": "SKU", "pronunciation": "skew", "caseSensitive": true}
  ]}'

This is a full replacement, not an addition. Pronunciations apply as an ordered list of whole-word rewrites, so entries can interact - editing one in isolation invites surprises. Read the list, change it, send it all back.

Brand skills

GET /v1/brand-kit/skills · POST /v1/brand-kit/skills PATCH /v1/brand-kit/skills/:id · DELETE /v1/brand-kit/skills/:id

A skill is markdown instructions - writing style, terminology, pacing, structure - folded into generation as a style guide. domain is videoScript or article.

curl -X POST https://api.vidocu.ai/v1/brand-kit/skills \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "SOP house style",
    "domain": "videoScript",
    "content": "## Voice\nSecond person, imperative. One action per step.\n\n## Terminology\nAlways \"work order\", never \"job\"."
  }'

Pass the returned id as scriptSkillId on analyze to apply it to a specific video. Content is capped at 20,000 characters.

Glossary

GET /v1/brand-kit/glossary · PUT /v1/brand-kit/glossary · POST /v1/brand-kit/glossary/import

Brand terms and how they should - or shouldn't - be translated. This feeds the translation pipeline directly, so it's the highest-leverage thing here if you translate.

Field	Meaning
`term`	The word or phrase
`doNotTranslate`	Keep it verbatim in every language
`translations`	Preferred rendering per language code; overrides `doNotTranslate` for that language
`caseSensitive`	Match case exactly
`notes`	Context for the model, e.g. "our product, not the body part"

curl -X PUT https://api.vidocu.ai/v1/brand-kit/glossary \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"entries": [
    {"term": "Pulse", "doNotTranslate": true, "notes": "our product, not a heart rate"},
    {"term": "work order", "translations": {"de": "Arbeitsauftrag", "fr": "ordre de travail"}}
  ]}'

Max 200 entries. PUT is a full replacement.

Importing from a spreadsheet

POST /v1/brand-kit/glossary/import

curl -X POST https://api.vidocu.ai/v1/brand-kit/glossary/import \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"csv": "term,doNotTranslate,de,fr\nPulse,true,,\nwork order,,Arbeitsauftrag,ordre de travail", "mode": "merge"}'

{ "mode": "merge", "added": 1, "updated": 1, "cleared": 0, "dropped": 0, "total": 34, "issues": [] }

merge (the default) updates matching terms and adds new ones. replace swaps the whole glossary.

Reading the counts: cleared is fields blanked because the CSV declared a column and left it empty - that's what makes export → edit in a spreadsheet → re-import work. dropped is the only signal that the 200-entry cap was hit. issues lists rows that were skipped and why, without failing the whole import.

Errors

Status	Code	Cause
400	`validation_error`	Nothing to update, or a value failed validation
401	`authentication_error`	Invalid or missing API key
403	`insufficient_scope`	Token lacks `brand_kit:read` / `brand_kit:write`
403	`insufficient_role`	The workspace role behind the token can't change settings
404	`not_found`	No such skill

Assets

GET /v1/brand-kit/assets · POST /v1/brand-kit/assets · DELETE /v1/brand-kit/assets/:id

The reusable files a workspace renders with: opening and closing scenes, watermarks, background music, backgrounds.

This endpoint records an asset; it does not carry the bytes. Upload the file with POST /v1/upload-url first, then register the resulting URL:

curl -X POST https://api.vidocu.ai/v1/brand-kit/assets \
  -H "Authorization: Bearer vdo_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "assetType": "opening_scene",
    "fileType": "video",
    "name": "Q3 intro",
    "url": "https://storage.vidocu.ai/..."
  }'

assetType is one of opening_scene, closing_scene, watermark, background_music, background, avatar_training. Filter the list by it.

Deleting removes the record. The stored file is left alone - other videos may still reference it, and removing it would break them.

Voices

Saved and cloned voices are served by GET /v1/voices, which returns the workspace's ready clones alongside the stock library, and DELETE /v1/voices/:id removes a clone. Creating a new clone needs an audio upload and isn't exposed yet.

Get the Brand Kit​

Update the Brand Kit​

Pronunciations​

Brand skills​

Glossary​

Importing from a spreadsheet​

Errors​

Assets​

Voices​