Aykroyd API Guide: Entity Detection to JSON Output
- Aykroyd turns text into JSON in under a second
- Free tier offers 5,000 characters per month
- It’s about 20% faster than comparable APIs
- Only English is supported, so non‑English docs need translation
What is the Aykroyd transformer encoder and how does it detect entities?
Aykroyd is a cloud‑based AI that turns raw documents into structured data. It works by first running a transformer encoder to spot entities, then applying a lightweight classifier that maps each entity to a predefined field. Launched on Sep 24, 2026, the service returns a JSON payload in under a second for a typical 500‑word input. The whole process happens on secure servers, so you never install heavy models locally. And because it’s offered via a REST API, any programming language can call it with a simple HTTP POST. So you can integrate it into existing workflows without redesigning your data pipeline.
Frequently asked questions
The Aykroyd transformer encoder is a neural network model that processes input text and uses attention mechanisms to identify and classify entities such as names, dates, and locations in a single pass.
Send an HTTP POST request to the Aykroyd endpoint with your text payload in the body. The API returns a JSON object containing detected entities, their types, and character offsets within milliseconds.
