Pay As You Go👌🏻
Turns code into Real-time Video Analyst
Apply object detection to your video stream or video file right away. Only pay for the execution time.
Try for free. No credit card required.
await flow.subscribe({
requestOptions: {
engine: "object_detection",
},
callbackSingleResult: async function (singleResult) {
// Excerept of fields:
singleResult.payload.forEach(function (p) {
p.trackingId;
p._class; // e.g. person
});
},
});
Power thousands of OpenCV applications
NPU Powered
Plug n' Play Video AI Analysis
Analyze video with AI without hassle. Focus on application development rather than setting up complicated infrastructure for running AI.
Serverless
We manage the top-notch NPU infrastructure for your AI model. You can focus on your application instead.
Super Compatible
Accepts different types of video stream (RTMP, RTSP, WebRTC etc.) and video file (mp4, mov, flv, etc.)
Fast Inference
Thanks to the optimized AI model. <50ms is needed to infer a 1080p video frame
Handful SDK
Getting result via SDK
We provide a dead simple SDK that fit any application. Just a function call is what you need.
Supported formats:
Coming soon:
NodeJS
Python
Free integration support
Set up
import rfc from "@socif/rfc_sdk";
const account = new rfc.Account({
userKey: "The API Key...",
});
const flow = await account.createFlow({
name: "My First Stream",
mode: "PULL",
pullUrl: "rtmp://example.com/live1",
});
Use
await flow.subscribe({
requestOptions: {
engine: "object_detection",
classes: ["person"],
},
callbackSingleResult: async function (singleResult) {
// You may access the following:
singleResult.inferenceTimestamp;
singleResult.payload.forEach(function (p) {
p.trackingId; // number
p._class; // enum string
p.bounds.x; p.bounds.y; p.bounds.w; p.bounds.h;
p.centroid.x; p.centroid.y;
});
},
});
Upload video stream
Inference results
Consume results in app
All covered
Supports Pull & Push
Whether you have the camera or the stream source, we get you covered without hassle.
Contact
Want to run a customised AI model? Contact us.
Testimonial
Reliable reviews from our customers
Trusted reviews from our partners demonstrate the effectiveness and user satisfaction of our service.
The solutions offered by your AI chatbots are truly impressive! We are able to communicate with our customers faster and more interactively.
Your chatbots offer a really great user experience! Thanks to its natural language processing capabilities, it can understand our users' questions and provide quick answers. Your AI chatbots have greatly streamlined our business' customer support process.
Thanks to the data analytics feature, we started to understand customer behavior better. Our chatbots provide us with information by analyzing user interactions.
The solutions offered by your AI chatbots are truly impressive! We are able to communicate with our customers faster and more interactively.
The solutions offered by your AI chatbots are truly impressive! We are able to communicate with our customers faster and more interactively, which increases customer satisfaction.
Contact
Want to reduce AI running cost by 80%? Contact us.
Pricing plans for you
Pay-As-You-Go
The cheapest way to get started.
- Video streams or files
- Supports 1080p@24FPS
- Volume Based Pricing
- Custom AI model
Reserve Plan
Great for 24/7 streaming.
- Video streams or files
- Supports 1080p@24FPS
- Unlimited Volume
- Custom AI model
e.g. 8 hours of stream per day
Benchmark for most of the use cases (Include Free Tier)
- RTSP
- 1080p@24 FPS
- Pay-As-You-Go
- 8 hours Free Tier
Simple Math
Cost sample
Free Tier: Users can enjoy 8 hours free quota every month
Frequently Asked Questions 💬
We answer your questions
Explore this section to learn more about our Real-time AI Inference and find answers to your questions.
How can I integrate AI Inference queue?
Integrating the inference queue into your system is a breeze! All you need to do is follow three simple steps:
1) Create queue with our CLI,
2) Install RF Cloud SDK in your application,
3) Consume the queue in your application. Voila!
See Docs for more details
How fast is your AI inference?
We have optimized the YOLOv8 for our cutting edge NPU architecture. For a 100KB frame image, the inference completes in <50ms!
How do your service help me to reduce cost by 80%?
Prepare to be amazed! Our hypervision technology allow us to unleash the potential of NPU servers. Our datacenter can infer the data with a cost of 80% less than GPU.
Time to move out from AWS, GCP, Azure.
What are the benefits of using the AI inference service from you?
Not only can you extract more value from the video, but also save your valuable time and cost to set up the infrastructure to run the AI model. Our top priority is to let the users focus on the most important thing.
What is the location of your datacenter?
Our datacenter is currently located in Hong Kong. We are planning to set up other datacenters in US, Europe and East Asia this year.
How is the data security of your service ensured?
Rest assured, the security of your data is our top priority. Our server will not store any video frame uploaded by the users. All images are destroyed right after processing.
Discord
Messenger
Feel free to contact us