Video Intelligence API
Build with video intelligence.
Integrate natural language video search, multimodal analysis, and structured outputs into your applications. One API for everything video.
search.py
import ceptory
# Search video with natural language
results = client.search(
query="person entering through back door",
filters={"confidence": 0.9}
)
for scene in results:
print(scene.timestamp, scene.summary)RESTful API
Standard HTTP endpoints with JSON responses
Enterprise Security
SOC 2 Type II, GDPR compliant, encrypted at rest
99.9% Uptime SLA
Enterprise-grade reliability and support
Multi-region
Deploy in US, EU, or private cloud
Usage Analytics
Real-time dashboards and monitoring
Team Management
Role-based access control and audit logs
API Capabilities
Everything you need for video intelligence
Six core endpoints that cover search, analysis, redaction, and export. Build complete video intelligence workflows with a single API.
Natural Language Search
Query video content using natural language. Find scenes, actions, objects, spoken phrases, and event sequences without manual tagging.
POST /v1/search
Multimodal Analysis
Extract structured insights from visual, audio, and speech signals. Scene classification, object detection, speaker diarization, and sentiment analysis.
POST /v1/analyze
Privacy Redaction
Automatically detect objects, activities, and sensitive operational events. Policy-aligned operational monitoring at scale.
POST /v1/redact
Background Replacement
Analyze scenes, speech, and timing. Perfect for operational review, safety compliance, and forensic investigation.
POST /v1/background
Structured Outputs
Get JSON-formatted results ready for downstream systems. Summaries, transcripts, timecoded events, and confidence scores.
POST /v1/export
Real-time Processing
Stream video for live analysis and alerts. Sub-second latency for security, compliance, and operational monitoring.
WS /v1/stream
Developer Experience
Ship faster with SDKs and examples
Official SDKs for Python, Node.js, Go, and Java. Comprehensive documentation, code examples, and a vibrant developer community.
import ceptory
client = ceptory.Client(api_key="your_api_key")
# Upload and index video
index = client.indexes.create(name="security-footage")
video = client.videos.upload(
index_id=index.id,
file="./footage.mp4"
)
# Natural language search
results = client.search(
index_id=index.id,
query="red sedan entering loading dock after midnight",
filters={"confidence": 0.85}
)
for result in results:
print(f"Found at {result.start_time}-{result.end_time}")
print(f"Summary: {result.summary}")Pricing
Simple, transparent pricing
Start with 10 free indexed minutes, then add prepaid balance for indexing and AI search usage.
Free
For evaluation and early workspace setup
Free
- 10 indexed video minutes included
- Developer dashboard and workspace setup
- Early search and retrieval evaluation
- Upgrade path into Stripe billing
Most popular
Developer
For production indexing and search workloads
Prepaid
- Affordable mode: $0.03 / minute
- Standard mode: $0.045 / minute
- Precision mode: $0.085 / minute
- Search API usage: $4 / 1000 queries
- Stripe prepaid workspace top-ups
Enterprise
For large-scale deployments and compliance
Custom
- Unlimited video processing
- Private cloud deployment
- Dedicated support
- Custom SLAs
- On-premise option
- Custom model training
FAQ
Frequently asked questions
What video formats does the API support?+−
Ceptory supports all major video formats including MP4, MOV, AVI, MKV, WebM, and more. We handle transcoding automatically so you can upload videos in any format and resolution.
How does natural language search work?+−
Our multimodal AI models analyze visual content, audio, speech, and on-screen text to create semantic embeddings. When you search, we match your query against these embeddings to find relevant scenes, actions, objects, and spoken phrases.
Can I deploy Ceptory in my own infrastructure?+−
Yes. Enterprise customers can deploy Ceptory in private cloud environments (AWS, GCP, Azure) or fully on-premise. This ensures data never leaves your security boundary.
What are the API rate limits?+−
Rate limits vary by plan. Starter tier allows 60 requests/minute. Pro and Enterprise tiers have higher limits and can be customized based on your needs.
How accurate is the face and license plate detection?+−
Our detection models achieve 99%+ accuracy on standard benchmarks. We continuously improve models and offer custom training for specific use cases like low-light footage or unusual camera angles.
Do you offer SDKs?+−
Yes. We provide official SDKs for Python, Node.js, Go, and Java. All SDKs include TypeScript definitions, automatic retries, and comprehensive documentation.
Get started
Ready to build with video intelligence?
Talk to our team to get API access, discuss your use case, or learn more about enterprise deployment options.
One API • All capabilities