01
Step 1
POST Video URL
Send a POST request with your video URL. Direct media links, HLS VOD manifests, or authenticated access are supported.
RESTful API + Webhooks
Video Description API for Developers
Integrate AI-powered video metadata generation into your platform. Submit video URLs, including HLS VOD manifests, via API and receive titles, descriptions, and tags via webhooks. Analyze both visual frames and audio content. Free sandbox for testing.
Step 1
POST Video URL
Step 2
Async Processing
Step 3
Receive Webhook
sample-video.mp4
00:02:34
vision_audio
frames:20
Request
# Create a job — vision, vision_audio, or audio mode
curl -X POST https://api.descrideo.com/v1/video-descriptions
-H "X-API-Key: YOUR_API_KEY"
-H "Content-Type: application/json"
-d '{"source":{"type":"direct_url"}}'
202 Accepted
{
"job_id": "01HXYZ123ABC",
"status": "queued",
"delivery": "webhook"
}
02
Step 2
Async Processing
API extracts frames and audio, runs AI analysis. Track status via job ID or wait for webhook.
03
Step 3
Receive Webhook
Get JSON with title, description, and tags via HMAC-signed webhook delivery.
Descrideo vs general-purpose tools
When your goal is reliable video metadata generation, purpose-built workflow beats stitching together generic multimodal APIs.
Descrideo
Recommended for metadata-first teams- Best for
- Async video description APIs, webhook workflows, and developer portals
- Strength
- Built specifically for titles, descriptions, tags, private video access, and webhook-first automation.
- Trade-offs
- Narrower scope than hyperscalers, but faster time-to-value for this exact use case.
Google Video AI
- Best for
- Large cloud ecosystems and custom media intelligence pipelines
- Strength
- Deep cloud integrations and broad ML tooling.
- Trade-offs
- More assembly work when all you need is production-ready metadata generation with webhooks.
Cloudinary
- Best for
- Media management, transformations, and delivery pipelines
- Strength
- Excellent asset delivery and optimization workflows.
- Trade-offs
- Video SEO metadata generation is usually one capability among many, not the core product surface.
AssemblyAI
- Best for
- Speech-heavy media and transcription-first products
- Strength
- Strong audio and transcription workflows.
- Trade-offs
- You may still need extra logic for frame analysis, combined outputs, and video-specific metadata formatting.
Why Developers Choose Descrideo API
Enterprise-grade video description API built for scale, security, and developer experience
Free Sandbox Testing
Test your integration with real webhook deliveries, no video processing required. Free daily quota included.
$ curl https://api.descrideo.com/v1/video-descriptions
-H "Authorization: Bearer nxt_test_key"
# webhook delivery without full production wiring
Async Webhook Delivery
Non-blocking API with HMAC-signed webhooks. Automatic retries with exponential backoff.
Portal API
queued
Webhook
signed
Your app
processed
Flexible Video Access
Direct media URLs, HLS VOD manifests, authenticated access, or Source Provider pattern. Keep your videos private.
direct_url
authenticated access
source_provider
Flexible Billing
Prepaid token balance or postpaid mode. Pay only for successful processing. No hidden fees.
Prepaid
Token balance
Postpaid
Pay on success
Audio Transcription
Three generation modes: vision-only, combined vision + audio, or audio-only. Leverage speech content for richer descriptions.
Enterprise Security
SSRF protection, data encryption at rest, HMAC-SHA256 signatures, and audit logging.
Signature
HMAC-SHA256
Logging
Audit trails
Storage
Private access
Simple REST API
Get started with just a few lines of code
# Create a job — vision, vision_audio, or audio mode
curl -X POST https://api.descrideo.com/v1/video-descriptions \
-H "X-API-Key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"source": {
"type": "direct_url",
"url": "https://example.com/video.mp4"
},
"frames": 20,
"generation_mode": "vision_audio",
"webhook": {"url": "https://your-app.com/webhooks/video"}
}'
# Response: 202 Accepted
{ "job_id": "01HXYZ123ABC", "status": "queued" }
Integration recipes for real teams
Use the same API from backend services, queues, CMS workflows, or low-code automation tools.
Python
PythonQueue a job from a FastAPI, Django, or Flask backend and forward results into your CMS.
import requests
payload = {
'source': {'type': 'direct_url', 'url': 'https://cdn.example.com/video/master.m3u8'},
'frames': 20,
'generation_mode': 'vision_audio',
'webhook': {'url': 'https://app.example.com/webhooks/descriptions'},
}
response = requests.post(
'https://api.descrideo.com/v1/video-descriptions',
headers={'Authorization': 'Bearer nxt_live_key'},
json=payload,
timeout=30,
)Node.js
Node.jsAttach metadata generation to upload completion events or background workers in your product stack.
const response = await fetch('https://api.descrideo.com/v1/video-descriptions', {
method: 'POST',
headers: {
Authorization: 'Bearer nxt_live_key',
'Content-Type': 'application/json',
},
body: JSON.stringify({
source: { type: 'direct_url', url: 'https://example.com/demo.mp4' },
frames: 30,
generation_mode: 'vision',
webhook: { url: 'https://app.example.com/webhooks/descriptions' },
}),
})PHP
PHPSend jobs from Laravel actions, queued jobs, or admin tools and store the returned job ID for reconciliation.
use Illuminate\Support\Facades\Http;
$response = Http::withToken('nxt_live_key')
->post('https://api.descrideo.com/v1/video-descriptions', [
'source' => ['type' => 'source_provider', 'provider' => ['asset_id' => 'video_123']],
'frames' => 20,
'generation_mode' => 'vision_audio',
'webhook' => ['url' => 'https://app.example.com/webhooks/descriptions'],
]);No-code
No-codeUse Make, Zapier, n8n, or an internal workflow runner to chain uploads, webhooks, and content publishing.
1. New video uploaded to storage
2. POST create-job request to Descrideo
3. Wait for signed webhook callback
4. Parse title, description, and tags
5. Push metadata into CMS, catalog, or search indexBuilt for Every Use Case
From video platforms to accessibility workflows, our API powers diverse applications
Pilot results from real workflows
These are representative internal and pilot-style implementations showing the kind of throughput and discoverability gains teams can expect.
Pilot case
01
UGC marketplace ingestion
Challenge
A content marketplace needed every uploaded video to ship with searchable titles and tags before editorial review.
Implementation
Descrideo was triggered after upload completion and returned metadata asynchronously into the moderation queue.
- Metadata coverage increased from 18% to 92% of newly uploaded videos.
- Average manual cataloging time dropped from roughly 2 business days to under 30 minutes.
- Editors started review with pre-filled multilingual titles, descriptions, and tags.
Accessibility workflow
02
Training library accessibility backlog
Challenge
An internal learning portal needed draft descriptions for hundreds of archived training recordings.
Implementation
Vision + audio mode generated first-pass summaries and structured notes for accessibility review teams.
- Reviewers cut first-pass drafting time by about 60%.
- Teams used one webhook payload to populate LMS summaries and accessibility checklists.
- Content owners gained a consistent metadata baseline across older recordings.
Catalog automation
03
Product video SEO at scale
Challenge
An ecommerce catalog needed consistent descriptions for demo clips, unboxings, and comparison videos.
Implementation
Descrideo generated SEO-ready copy that merchandising teams reviewed before publishing to product pages.
- Publishing teams reduced copy creation time from hours to minutes per launch batch.
- Search teams received structured tags and VideoObject-ready fields for product pages.
- Merchandisers kept brand tone by attaching custom instructions to each batch.