API Documentation

Complete reference for the Viffly.app REST API

Contents

Authentication Credits AI Video (Alibaba & MiniMax)Video thumbnails Webhooks n8n Integration Reel generation Videos AI Services API Keys Media Errors

Authentication

The API supports two authentication methods. Both require aBusinessplan or higher for external calls.

Method 1: API key (recommended)

curl -X POST https://viffly.app/api/generate-sync \
  -H "X-API-Key: vf_votre_cle_api" \
  -F "audio=@narration.mp3" \
  -F "title=Ma vidéo"

Method 2: Bearer Token

# Obtenir un token via /api/login
# session : "short" (12 h) | "default" (7 j) | "long" (30 j)
curl -X POST https://viffly.app/api/login \
  -H "Content-Type: application/json" \
  -d '{"email":"vous@email.com","password":"...","session":"short"}'

# Réponse : { "token": "eyJhbG...", "expires_at": "...", "session_ttl": "12h" }

# Utiliser le token
curl -H "Authorization: Bearer eyJhbG..." \
  https://viffly.app/api/me

Bearer Tokens expire after12 h, 7 d or 30 d depending on the `session` parameter. API keys (vf_...) are scoped, revocable, rotatable and rate-limited individually — this is the method to use in production.

Key permissions (scopes)

Every key carries a list of scopes. A call to an endpoint it does not cover is rejected with a 403 naming the missing scope — grant the minimum needed.

Scope	Endpoints covered
generate	/api/ltx/* · /api/generate · /api/generate-sync · talking-head · restyle · montage · longshort · loop-assemble · director · slideshow · reframe · autocut · audio-clean · watermark · voiceover · media-jobs
ai_images	/api/image-jobs · /api/generate-images · /api/images/edit · /api/characters/*
ai_script	/api/ai/*
videos	/api/videos
credits	/api/me · /api/credits/*

A key with no stored scopes (created before scopes existed) passes everywhere. Rotate it to apply a permission set.

Key lifecycle

Up to 5 keys per account. The full secret is shown only at creation — it can never be displayed again.
Per-key rate limit (60 req/min by default, 1000 max). Exceeding it returns 429 with a Retry-After header.
Expiry of your choice: 7, 30, 90, 365 days or never. An expired key is rejected without being deleted.
Rotation: a new secret replaces the old one instantly, keeping name, scopes and limits. Use it the moment a key may have leaked.

# Renouveler le secret d'une clé (nom, scopes et limites conservés)
curl -X POST https://viffly.app/api/keys/{id}/rotate \
  -H "Authorization: Bearer eyJhbG..."

# Déconnecter tous les autres appareils
curl -X POST https://viffly.app/api/account/revoke-sessions \
  -H "Authorization: Bearer eyJhbG..."

Credit system

Each AI operation consumes credits (fixed cost for standard features, per-duration for AI video). Free operations (Pexels, Whisper) consume nothing.

🎬

Video render

5 cr.

📝

AI script

1 cr.

🎙️

AI voice

4 cr.

🖼️

AI image (×1)

2 cr.

🎵

AI music

5 cr.

💡

Image prompts

Free

📷

Pexels images

Free

💬

Transcription

Free

Examples:Budget video (your audio + Pexels) =5 cr.| Standard video (script + AI voice) =10 cr.| Full-AI video (fully automatic) =21 cr.

GET/api/credits/costs

Returns the credit cost of each operation, the available plans and usage scenarios.

Response

{
  "costs": { "render": 5, "ai_script": 1, "ai_voice": 4, "ai_image": 2, "ai_music": 5, "pexels": 0 },
  "packages": [
    { "slug": "free", "label": "Free", "monthlyCredits": 15 },
    { "slug": "pro", "label": "Pro", "monthlyCredits": 200 }, ...
  ],
  "scenarios": [ { "name": "Budget", "credits": 5 }, ... ]
}

GET/api/meAuth

Returns the account information, including the credit balance.

Response

{
  "id": "abc123",
  "client_email": "vous@email.com",
  "package": "pro",
  "creditsTotal": 142,
  "creditsMonthly": 130,
  "creditsPack": 12,
  "monthlyAllowance": 150,
  "packageLabel": "Pro",
  "creditCosts": { "render": 5, "ai_voice": 4, ... }
}

AI video generation (MiniMax / Alibaba)

Generates a video fromtext(t2v) or from animage(i2v) via the AI models. This is the API to use from n8n. The flow isasynchronous: you submit a job, then poll its status untildone.

🌊 Hailuo 2.3 (MiniMax)

MiniMax Hailuo 2.3 · model: "hailuo23"

🎬 Wan 2.2

Alibaba Wan 2.7 · model: "wan27"

Requires aProplan or higher. Authentication viaX-API-Key.

GET/api/ltx/pricing

Lists the tiers (resolution × duration), aspect ratios and scenarios. No auth required — useful to populate a UI or pick a tier_id.

Response

{
  "tiers": [
    { "id": "tier1", "label": "720p · 5s",  "resolution": "720p",  "seconds": 5,  "credits": 10 },
    { "id": "tier2", "label": "720p · 5s",  "resolution": "720p",  "seconds": 5,  "credits": 10 },
    { "id": "tier3", "label": "720p · 10s", "resolution": "720p",  "seconds": 10, "credits": 20 },
    { "id": "tier6", "label": "1080p · 5s", "resolution": "1080p", "seconds": 5,  "credits": 10 },
    { "id": "tier7", "label": "1080p · 10s","resolution": "1080p", "seconds": 10, "credits": 20 }
  ],
  "models": [
    { "id": "wan27", "provider": "dashscope_wan", "tiers": [...] },
    { "id": "hailuo23", "provider": "minimax_hailuo", "tiers": [...] }
  ],
  "aspects": [ { "id": "9:16" }, { "id": "1:1" }, { "id": "16:9" } ],
  "scenarios": [ { "id": "cinematic" }, { "id": "product" }, ... ],
  "default_negative": "ugly, blurry, low quality, ..."
}

POST/api/ltx/generate10 cr.Auth

Submits a video generation job. Returns a job_id immediately (non-blocking). JSON body.

Request body

Field	Type	Description
mode*	string	"t2v" (text→video) or "i2v" (image→video)
prompt*	string	Video description (≥ 3 characters)
tier_id	string	tier1, tier2, tier3, tier6, or tier7. Check /pricing for engine-specific combinations.
model	string	"hailuo23" (MiniMax), "happyhorse" or "wan27" (Alibaba)
image_base64	string	Starting image in base64 (required if mode=i2v). Without the data: prefix
image_url	string	Alternative: public HTTPS URL of the starting image (i2v)
negative	string	Negative prompt (default: see /api/ltx/pricing)
aspect	string	"9:16" (default), "1:1", "16:9"
scenario_type	string	Creative direction (see /api/ltx/pricing)
fps	number	Frames/second (default: 24)
seed	number	Seed (reproducibility). Random if omitted
audio_on	boolean	Keep the generated audio track (default: true)

Response

{
  "job_id": "915c93ea-58b7-4808-a83f-5dfa98ee1ef5",
  "status": "processing",
  "tier": { "id": "tier2", "label": "720p · 5s", "credits": 10 }
}

Example — text → video:

curl -X POST https://viffly.app/api/ltx/generate \
  -H "X-API-Key: vf_votre_cle" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "t2v",
    "model": "hailuo23",
    "prompt": "a sports car drifting on a coastal road at sunset, cinematic",
    "tier_id": "tier2",
    "aspect": "9:16"
  }'

Example — image → video:

curl -X POST https://viffly.app/api/ltx/generate \
  -H "X-API-Key: vf_votre_cle" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "i2v",
    "model": "wan27",
    "prompt": "slow cinematic push in, gentle motion",
    "tier_id": "tier2",
    "image_base64": "'"$(base64 -w0 photo.jpg)"'"
  }'

POST/api/ltx/generate-sync10 cr.Auth

SYNCHRONOUS version: submits, waits for the render internally, then returns the video — in ONE single call (no wait/poll loop). Ideal for a single n8n node. Set a long timeout (e.g. 1200000 ms).

Request body

Field	Type	Description
(same as /api/ltx/generate)*	json	mode, prompt, tier_id, model, image_base64/image_url, aspect…

Response

// Par défaut → renvoie le FICHIER MP4 (binaire). Dans n8n : Response Format = File.
// Avec ?format=json → renvoie l'URL à la place :
{
  "status": "done",
  "job_id": "915c93ea-...",
  "output_url": "/videos/ltx_915c93ea-....mp4",
  "url": "https://viffly.app/videos/ltx_915c93ea-....mp4"
}

# Récupère directement le MP4 (sortie binaire)
curl -X POST https://viffly.app/api/ltx/generate-sync \
  -H "X-API-Key: vf_votre_cle" -H "Content-Type: application/json" \
  -d '{"mode":"t2v","model":"hailuo23","prompt":"...","tier_id":"tier2"}' \
  --max-time 1200 -o video.mp4

# Ou récupère juste l'URL (JSON)
curl -X POST "https://viffly.app/api/ltx/generate-sync?format=json" \
  -H "X-API-Key: vf_votre_cle" -H "Content-Type: application/json" \
  -d '{"mode":"t2v","prompt":"...","tier_id":"tier2"}'

GET/api/ltx/status/:idAuth

(Async mode) Polls the state of a job. Call it in a loop (every ~5-10 s) until status = done or error. Not needed if you use /generate-sync.

Response

// en cours
{ "id": "915c93ea-...", "status": "processing", "progress": 50 }

// terminé
{
  "id": "915c93ea-...",
  "status": "done",
  "output_url": "/videos/ltx_915c93ea-....mp4",  // relatif → préfixer https://viffly.app
  "progress": 100
}

// échec (crédits remboursés)
{ "id": "915c93ea-...", "status": "error", "error": "..." }

output_urlisrelative. Full download URL =https://viffly.app+output_url.

Smart thumbnails

Turn landscape, vertical, or square video into multi-format thumbnails. Choose real frames, AI-retouched frames, or new visuals generated from the topic and transcript.

16:9 · 9:16 · 1:1 · 4:5Smart crop for each destination

1–6 × formatGenuinely different moments

Async · n8n readyDurable jobs with progress tracking

POST/api/thumbnails5 cr.Auth

Start a job from a video URL, authorized S3 source, or a file already hosted by Viffly.

Request body

Field	Type	Description
source_url*	string	URL HTTPS publique, s3:// autorisé ou /videos/...
count	number	1 à 6 variantes par format (défaut : 3)
formats	string[]	auto, 16:9, 9:16, 1:1, 4:5
visual_mode	string	frames (défaut), enhance ou generate
ai_quality	string	standard (2 cr./image) ou premium (5 cr./image)
image_prompt	string	Direction facultative pour la retouche ou la génération
language	string	Langue du titre automatique : auto, fr, en, ar, es, de ou pt
title	string	Titre optionnel ; généré depuis le contenu si omis
subtitle	string	Sous-titre optionnel
style	string	editorial, bold, clean, minimal
tone	string	auto, warm, cool, mono
smart	boolean	Analyse visuelle + transcription + point focal
include_text	boolean	Ajouter ou non la composition typographique
accent	string	Couleur hexadécimale, ex. #6D4AFF

Response

{
  "job_id": "53a9c34d-...",
  "status": "processing",
  "cost": 11,
  "visual_mode": "generate",
  "cost_breakdown": { "composition": 5, "ai_images": 6 },
  "outputs_planned": 3,
  "status_url": "/api/thumbnails/53a9c34d-..."
}

curl -X POST https://viffly.app/api/thumbnails \
  -H "X-API-Key: vf_votre_cle" \
  -H "Content-Type: application/json" \
  -d '{
    "source_url": "https://cdn.exemple.com/video.mp4",
    "count": 3,
    "formats": ["16:9", "9:16"],
    "visual_mode": "generate",
    "ai_quality": "standard",
    "language": "fr",
    "image_prompt": "documentary editorial visual, authentic texture, no text",
    "style": "editorial",
    "smart": true
  }'

POST/api/thumbnails/upload5 cr.Auth

Convenient multipart upload for n8n and files up to 200 MB.

Request body

Field	Type	Description
video*	file	MP4, MOV, M4V, MKV, WebM, AVI ou MPEG · 200 Mo max
formats	string	Tableau JSON ou liste séparée par des virgules
count	number	1 à 6 variantes par format
visual_mode	string	frames, enhance ou generate
ai_quality	string	standard ou premium
image_prompt	string	Direction facultative pour l’IA
language	string	Langue du titre automatique : auto, fr, en, ar, es, de ou pt
style	string	editorial, bold, clean, minimal

Response

{ "job_id": "53a9c34d-...", "status": "processing", "status_url": "/api/thumbnails/53a9c34d-..." }

curl -X POST https://viffly.app/api/thumbnails/upload \
  -H "X-API-Key: vf_votre_cle" \
  -F "video=@episode.mp4" \
  -F 'formats=["16:9","9:16"]' \
  -F "count=3" \
  -F "visual_mode=enhance" \
  -F "ai_quality=standard" \
  -F "language=fr" \
  -F "image_prompt=cinematic light, cleaner background" \
  -F "style=editorial"

GET/api/thumbnails/:idAuth

Track the job and retrieve analysis, source frames, and every final thumbnail.

Response

{
  "job_id": "53a9c34d-...",
  "status": "done",
  "progress": 100,
  "analysis": { "summary": "..." },
  "visual_mode": "enhance",
  "outputs": [
    { "url": "/videos/thumb_....jpg", "format": "16:9", "source_kind": "enhanced", "master_url": "/videos/mmedit_....png", "raw_frame_url": "/videos/thumb_..._frame_1.jpg", "width": 1280, "height": 720, "timestamp": 42.1, "score": 91 }
  ]
}

n8n

Recommended chain: HTTP Request POST → Wait 3 s → HTTP Request GET → IF status = done. Loop back to Wait while status is processing.

GET https://viffly.app/api/thumbnails/{{ $('Viffly · Miniatures').item.json.job_id }}

Webhooks — get notified when a render finishes

Add callback_url to any job creation call: instead of polling for status, Viffly POSTs to you the moment the render completes. Your n8n or Make execution never sits blocked for minutes.

POST /api/ltx/generate
{
  "mode": "i2v",
  "prompt": "…",
  "callback_url": "https://votre-domaine.com/viffly-hook"
}

The address must be HTTPS and publicly reachable. Private addresses, localhost and plain HTTP are rejected at creation time, before any credits are charged.

What you receive

{
  "event": "job.completed",      // ou "job.failed"
  "job_id": "3f2a…",
  "kind": "video",
  "status": "done",              // ou "error"
  "output_url": "/videos/ltx_3f2a….mp4",
  "error": null,
  "delivered_at": "2026-07-27T19:44:12.031Z"
}

Headers

X-Viffly-Eventevent name, same as the event field in the body

X-Viffly-Deliveryunique delivery id — use it to discard duplicates

X-Viffly-Timestampunix timestamp in seconds, part of the signature computation

X-Viffly-Signaturesha256=… — HMAC of the body, lets you verify the call really comes from us

Verifying the signature

Compute the HMAC-SHA256 of "timestamp.body" with your secret and compare it to the header. Use the RAW request body: re-serialising the JSON changes whitespace and breaks the comparison.

const crypto = require('crypto');

function verifie(req, secret) {
  const ts   = req.headers['x-viffly-timestamp'];
  const recu = req.headers['x-viffly-signature'];
  const attendu = 'sha256=' + crypto
    .createHmac('sha256', secret)
    .update(ts + '.' + req.rawBody)   // corps BRUT, non re-sérialisé
    .digest('hex');
  return crypto.timingSafeEqual(Buffer.from(recu), Buffer.from(attendu));
}

On failure, up to 4 spaced attempts (immediate, +2s, +15s, +60s). A 4xx response is treated as a definitive refusal and is not retried. Return 2xx on receipt.

n8n integration — HTTP Request node

The video flow is asynchronous. In n8n, the pattern is:Submit→Wait→Check the status→ loop while ≠done.

① HTTP Request (POST /generate)② Wait 10s③ HTTP Request (GET /status)④ IF status = done?→ no: back to ②→ yes: download

n8n node builder

Configure the options below then generate the HTTP Request node JSON — copy it and paste it directly into the n8n canvas (Ctrl+V), or download the file.

Node typeAPI key"API" tab of the dashboard

ModeModelAspect ratio

Tier (resolution × duration)Prompt source

FPSSeed (optional)Audio

Negative prompt (optional)

Or configure the nodes manually:

NODE 1Submit the job — HTTP Request

Method	POST
URL	https://viffly.app/api/ltx/generate
Authentication	None (we pass the key in the header below)
Send Headers	ON → Name: X-API-Key \| Value: vf_votre_cle
Send Body	ON → Body Content Type: JSON
Specify Body	Using JSON

Body (JSON)

{
  "mode": "i2v",
  "model": "hailuo23",
  "prompt": "{{ $json.prompt }}",
  "tier_id": "tier2",
  "aspect": "9:16",
  "image_base64": "{{ $binary.data.data }}"
}

i2v: if the image comes from a previous node (e.g. Telegram, Read Binary File), reference its base64 binary data with{{ $binary.data.data }}. For t2v, removeimage_base64and set"mode": "t2v".

NODE 2 + 3Wait then check the status

Node 2 — Wait: 10 seconds.Node 3 — HTTP Request(poll):

Method	GET
URL	https://viffly.app/api/ltx/status/{{ $('NODE 1').item.json.job_id }}
Send Headers	ON → X-API-Key: vf_votre_cle

NODE 4Loop — IF status = done?

NodeIF— condition (String):{{ $json.status }} equals done.

false(not ready yet) → connect the output toNode 2 (Wait): the loop continues.
true(done) → continue. The final URL =https://viffly.app{{ $json.output_url }}

Download the video (HTTP Request)

Method	GET
URL	https://viffly.app{{ $json.output_url }}
Response Format	File / Binary

n8n tips

To avoid an infinite loop: add a counter (Set node + IF on a max number of iterations, e.g. 60 ≈ 10 min).
Wan 2.2 takes 3-5 min: set the Wait to 15-20 s to reduce the number of calls.
The cost is charged once at launch andautomatically refundedif the render fails.
No incoming webhook: it's polling. An API keyvf_…("API" tab) is enough.

Reel generation (audio + images)

POST/api/generate-sync5 cr.Auth

Generates a video and waits for the result (20 min timeout). Returns the download URL.

Request body

Field	Type	Description
audio*	file	MP3 audio file of the narration
images	file[]	Images (1 to 4 files). Alternative: image_url
image_url	string	URL of an image (alternative to images)
image_url_0..N	string	Indexed image URLs, no longer capped at 3 (up to 30 media). An .mp4/.mov URL here plays as a video clip at that position.
media	string[]	Ordered JSON array mixing images AND videos — array order = display order. Videos detected by extension (.mp4, .mov, .webm…).
format	string	Output format: 9:16 (default), 16:9, 1:1, 4:5, 4:3
title	string	Title shown in the intro
source	string	Source / site name
script_text	string	Script text for the subtitles
template	string	32 templates : default, bold, clean, news, minimal, highlight, cinematic, neon, retro, gaming, story, meme, comic, glitch, noir, vaporwave, duotone, matrix, ticker…
caption_style	string	karaoke, glow, outline, neon, block, shadow, gradient, typewriter
caption_size	string	S, M, L, XL
caption_position	string	top, center, bottom
language	string	fr, en, ar, es, de, it, pt
profile	string	Render profile name (e.g. shadesplays)
music_url	string	Background music URL (loops automatically when the video outlasts the track)
music_start_at	number	Music start offset (seconds)
voice_type	string	"ai" for MiniMax narration (+4 cr.)
voice_id	string	MiniMax voice ID
outro_text	string	End screen text
outro_sub	string	End screen subtitle

Response

{
  "videoId": "7865aac1-c301-4060-bd6e-6ca4fe3f6859",
  "url": "https://s3.eu-west-3.amazonaws.com/.../render.mp4",
  "thumbnail": "https://s3.eu-west-3.amazonaws.com/.../thumb.jpg",
  "duration": 27.5,
  "size": 8542190
}

Full example:

curl -X POST https://viffly.app/api/generate-sync \
  -H "X-API-Key: vf_votre_cle" \
  -F "audio=@narration.mp3" \
  -F "format=16:9" \
  -F 'media=["https://cdn.exemple.com/clip.mp4","https://cdn.exemple.com/photo1.jpg","https://cdn.exemple.com/photo2.jpg"]' \
  -F "title=Mon titre" \
  -F "source=MonSite.com" \
  -F "template=gaming" \
  -F "caption_style=karaoke" \
  -F "caption_size=L" \
  -F "script_text=Voici le texte des sous-titres" \
  -F "profile=shadesplays"

POST/api/generate5 cr.Auth

Starts the generation in the background. Returns a jobId to track progress via SSE.

Response

{ "jobId": "e9e2d4f9-..." }

Track progress:GET /api/videos/events?jobId=...(Server-Sent Events)

Videos

GET/api/videosAuth

Lists all the account's videos, sorted by descending date.

Response

[
  {
    "id": "7865aac1-...",
    "title": "Mon titre",
    "status": "done",
    "url": "https://...",
    "thumbnail": "https://...",
    "size": 8542190,
    "duration": 27.5,
    "created": "2026-03-28T..."
  }, ...
]

GET/api/videos/:id.mp4Auth

Download a video's MP4 file.

DELETE/api/videos/:idAuth

Delete a video and its associated files.

AI services

POST/api/ai/script1 cr.Auth

Generates a narration script optimized for virality.

Request body

Field	Type	Description
prompt*	string	Topic or instruction
existing	string	Existing script to improve
context	string	Additional context (web content)
language	string	Target language (fr, en, ar...)

Response

{ "script": "Saviez-vous que...", "charsUsed": 342, "charsMax": 700 }

POST/api/generate-images2 cr.Auth

Generates images via FAL Flux 2 Pro.

Request body

Field	Type	Description
prompt*	string	Image description
count	number	Number of images (1-3, default: 3)

Response

{ "images": ["https://fal.media/...", ...] }

Cost: 2 credits × number of images

POST/api/ai/image-promptAuth

Generates 3 image prompts optimized for a given topic.

Request body

Field	Type	Description
topic*	string	Video topic
style	string	Desired visual style

Response

{ "prompts": ["A cinematic...", "An aerial...", "A close-up..."] }

POST/api/generate-music5 cr.Auth

Generates background music via MiniMax Music.

Request body

Field	Type	Description
prompt	string	Description of the musical style
style	string	Music genre (pop, electronic, cinematic...)
instrumental	boolean	Instrumental only (default: true)
title	string	Track title

Response

{ "taskId": "abc123", "status": "pending" }

Check the status:GET /api/generate-music/:taskId

GET/api/voicesAuth

Lists the available MiniMax voices (1h cache).

Response

{ "voices": [{ "voice_id": "pNInz6ob...", "name": "Adam", "labels": {...} }, ...] }

API key management

GET/api/keysAuth

Lists all your API keys.

Response

[{
  "id": "uuid", "name": "Production", "key_prefix": "vf_a1b2c3d4",
  "scopes": ["generate","videos"], "rate_limit": 60,
  "expires_at": null, "last_used_at": "2026-03-28T...",
  "usage_count": 42, "is_active": true
}]

POST/api/keysAuth

Creates a new API key. The full key is only returned at creation.

Request body

Field	Type	Description
name	string	Key name (e.g. Production)
scopes	string[]	Permissions: generate, videos, credits, ai_script, ai_images, ai_music
rate_limit	number	Requests per minute (1-1000, default: 60)
expires_in_days	number	Days before expiration (null = never)

Response

{
  "id": "uuid",
  "key": "vf_a1b2c3d4e5f6...",  // ⚠️ Affiché UNE SEULE FOIS
  "name": "Production",
  "scopes": ["generate", "videos", "credits"]
}

PATCH/api/keys/:idAuth

Edit a key (name, permissions, rate limit, enable/disable).

Request body

Field	Type	Description
name	string	New name
is_active	boolean	Enable/disable
scopes	string[]	New permissions
rate_limit	number	New rate limit

DELETE/api/keys/:idAuth

Permanently deletes an API key.

Media library

GET/api/musicAuth

Lists the music library (uploaded tracks + MiniMax).

GET/api/pexels/imagesAuth

Pexels image search.

Request body

Field	Type	Description
q	query	Search term
page	query	Page (default: 1)
per_page	query	Results per page (default: 20)

GET/api/pexels/videosAuth

Pexels video search.

Error codes

Code	Meaning	Action
401	Invalid or expired token	Renew the token or create a new API key
402	Insufficient credits	Buy a pack or upgrade to a higher plan
403	Insufficient plan (Business+ required)	Upgrade to Business or Agency
429	Rate limit exceeded	Wait 1 minute or increase the rate limit
500	Server error	Retry or contact support

Standard error format

{
  "error": "Message descriptif de l'erreur",
  "creditsNeeded": 5  // (optionnel) crédits nécessaires
}

Rate limiting

The default limits are:

API keys: configurable per key (1-1000 req/min, default 60)
Bearer Token: 100 req/min global
AI endpoints (script, images, music): 10 req/min
Video generation: 3 concurrent max per account