Video Intelligence API
Raw extraction is table stakes. VidContext adds an intelligence layer — expert scoring frameworks that evaluate video effectiveness, quality, and strategy across 8 specialized modes. Get numerical scores, detailed assessments, and actionable recommendations from a single API call.
From extraction to intelligence
Most video APIs stop at extraction: here is what was said, here is what appeared on screen, here are the detected objects. That data is useful, but it does not answer the questions that matter. Is this ad effective? Will this product video convert? Is this creator producing quality content? Is this UGC safe for our brand?
VidContext answers those questions directly. Each analysis mode applies 6 expert-designed scoring frameworks that evaluate the video like a specialist would — returning scores, reasoning, and specific recommendations. The difference between a video description API and a video intelligence API is the difference between data and insight.
8 modes, 42 scoring frameworks
Ad Analysis
Score advertising effectiveness: hook strength, CTA clarity, emotional resonance, brand integration.
Creator Analysis
Evaluate content creator performance: engagement tactics, audience retention signals, production quality.
E-commerce
Assess product video conversion potential: benefit clarity, objection handling, purchase motivation.
Training
Rate instructional content: learning structure, concept clarity, pacing, knowledge retention design.
UGC Vetting
Screen user-generated content: brand safety, authenticity signals, quality thresholds, compliance flags.
Competitor Intel
Decode competitor strategy: messaging positioning, production investment, audience targeting, content patterns.
All modes also include the base Context extraction: transcript, visual scenes, on-screen text, brands, and audio analysis.
Scored analysis in one request
Switch the output_format parameter to any analysis mode. The response includes all base extraction data plus structured scores and recommendations.
curl -X POST https://api.vidcontext.com/v1/analyze \ -H "X-API-Key: vc_your_key" \ -F "source=https://example.com/facebook-ad.mp4" \ -F "mode=ad"
{
"metadata": { "duration": "0:32", "resolution": "1080x1080" },
"transcript": [{ "time": "0:00-0:05", "text": "Tired of editing videos manually?" }],
"scoring": {
"hook_strength": {
"score": 8,
"assessment": "Opens with a direct pain point question in first 2 seconds",
"recommendation": "Add visual pattern interrupt to complement the verbal hook"
},
"cta_effectiveness": {
"score": 6,
"assessment": "CTA appears at 0:28 but competes with closing animation",
"recommendation": "Isolate CTA in final 3 seconds with clean background"
},
"brand_integration": {
"score": 7,
"assessment": "Logo visible throughout, product demo is the core content",
"recommendation": "Add brand name in verbal sign-off for audio-only viewers"
}
},
"overall_score": 7.2,
"top_recommendations": [
"Strengthen CTA isolation in final seconds",
"Add captions — 85% of social video is watched on mute"
]
}What teams use video intelligence for
- Ad performance audits — Score every ad in your creative library against effectiveness frameworks. Identify which ads work and why, without subjective reviews.
- Competitor video analysis — Decode competitor ad strategy, content patterns, and production quality at scale. Track changes over time.
- Creator vetting at scale — Evaluate influencer content quality, brand safety, and audience engagement signals before signing partnerships.
- E-commerce listing optimization — Score product videos against conversion frameworks and get specific recommendations to improve purchase rates.
Frequently asked questions
What makes this different from video description or tagging?
Video description and tagging tell you what is in the video. Video intelligence tells you what it means. VidContext applies expert scoring frameworks that evaluate effectiveness, quality, and strategy — returning numerical scores, rankings, and specific recommendations, not just labels.
What are the 8 analysis modes?
Context (general extraction), Ad Analysis (advertising effectiveness scoring), Creator Analysis (content creator performance evaluation), E-commerce (product video conversion optimization), Training (instructional content quality), UGC Vetting (user-generated content brand safety), and Competitor Intelligence (competitive video strategy analysis).
How are scores calculated?
Each analysis mode applies 6 specialized scoring frameworks. For example, Ad Analysis evaluates hook strength, messaging clarity, emotional resonance, call-to-action effectiveness, brand integration, and production quality. Each framework returns a numerical score with detailed reasoning and improvement recommendations.
Can I use custom scoring criteria?
Currently, VidContext provides 42 built-in frameworks across 8 modes, designed by marketing and content experts. Custom scoring criteria are on the roadmap. In the meantime, you can use the Context mode to extract raw data and apply your own scoring logic on top.
What does a typical analysis output look like?
A scored analysis returns the same base data as Context mode (transcript, scenes, text, brands, audio) plus a structured scoring section. Each framework includes a numerical score (1-10), detailed assessment, strengths, weaknesses, and specific recommendations for improvement.
Ready to start?
5 free analyses without an account. 20 credits on signup. No credit card required.
Try free