DirectAI
Create and deploy computer vision models
About DirectAI
DirectAI allows anyone to create and deploy computer vision models in seconds, without training data or code. DirectAI is built on top of Vision Language Models (VLMs) that have already been trained on billions of image-caption pairs. This enables the creation (and even modification) of models without any image data.
Adding detection to your images, videos, or streams only requires one API call. DirectAI caters specifically to Ant Media customers who are currently unable to make use of object data present in their streams (e.g. identify defects, count people, continuously map a scene, etc.).
Who can use this plugin?
DirectAI provides comprehensive object recognition solutions that cater to the diverse needs of video surveillance products across various industries, including construction, retail, agriculture, and entertainment.
How the plugin works?
Getting Started
Before Starting:
2. You need to generate DirectAI client credentials by signing up here and schedule an onboarding session with the founders.
Deploying:
During the onboarding call, the team will help you clone and deploy the DirectAI x Ant Media plugin inside your AMS.
Using:
Once you’ve got everything working end-to-end, DirectAI enables a massive amount of customization.
- Upload individual frames to AWS S3 for future use.
- Modify your object tracking model configuration with plain language.
- Track multiple objects and types of objects simultaneously.
- Send detection/tracking results directly to a webhook.
- Start and stop tracking via REST API calls to your AMS.
Support
Please contact ben@directai.io or schedule a meeting with the founding team for custom help. DirectAI’s discord has an active community that is ready to collaborate and answer questions.
