📘
Prerequisites

✨ Create a Workspace Set up your environment in the HrFlow.ai portal.

🔑 Get your API Key Retrieve your API_KEY from the settings.

🧠 Activate Job Embedding API Ensure the embedding service is enabled for your plan.

The Embedding API generates SOTA, HR-native, and cross-lingual vectors for Jobs to power similarity search, retrieval, clustering, and HR-grade ML. Unlike generic LLM vectors, these encoders are pretrained on 1.2B hiring decisions to capture signals beyond keywords to identify "Intent-Similar" jobs a candidate is likely to click and apply to.

This endpoint allows you to vectorize jobs by providing either raw data objects or references to profiles already indexed in your sources. Results are deterministic: the same inputs will consistently return the same vectors.

1. How to get a Job Object

This endpoint requires structured data as input. You can obtain a valid HrFlow.ai Job Object through two primary methods:

Manual Construction: Build the object programmatically by following our Job Object Specification 📖 The Job Object .
Automated Parsing: Convert a raw job description or file into a structured Job Object using our Parsing API (e.g. Job Parsing). This is an ease-of-use feature that ensures your job data is normalized and "HR-ready" for vectorization.

2. Available Algorithms

Choose the right algorithm key based on your use case, data type, and precision needs.

Algorithm Keys	Use Case	Status
`embedding-hyperion-1-jobs-1024d`	Best for Job ↔ Job similarity and clustering based on role context.	🟢 Live
`embedding-hyperion-1-jobs-128d`	High-speed version for large-scale clustering and deduplication.	🟢 Live
`embedding-hyperion-1-dual-1024d`	Matching and Retrieval. Represents jobs and profiles in a shared space.	🟢 Live
`embedding-hyperion-1-cross-1024d`	Optimized for Profile ↔ Job fit scoring using pairwise representation.	⚪ Coming Soon

3. Request headers

Field	Value
`Content-Type`	`application/json`
`X-API-KEY`	Your API Key.
`X-USER-EMAIL`	Your HrFlow.ai account's email.