Machine reading comprehension parses requirements, perks and company info from any JD into standard role tags.
Parses job descriptions in any format.
Produces multi-dimensional role tags.
Aligns with candidate profile tags.
Surfaces deeper role signals.
Auto-classifies required skills.
Parses 50 job posts per second.
Deep-learning sequence labeling and NER, paired with active learning, reliably extract requirements, perks and company info from unstructured JD text.
Read the guide →Send the JD text to the API and get standard role tags back in seconds. Python, JavaScript, Ruby, cURL.
Extract 80+ standard tags — company, title, years, degree, skills, salary, location, benefits — from one JD.
Role tags share one taxonomy with candidate profiles — feeding matching directly, aligned at the semantic level.
Standard role profiles plug into your ATS, HR SaaS and enterprise hiring systems to power search and matching.
Infrastructure and data services run on Alibaba Cloud — physical-machine encryption, vulnerability scanning and web-attack defense.
All traffic is encrypted with 256-bit SSL/HTTPS certificates, so nothing is leaked or tampered with in transit.
Private data uploaded to our servers is auto-deleted on a daily schedule, minimizing unforeseen data risk.
A randomized algorithm encrypts every resume during upload and parsing — even our own staff cannot read the plaintext.
One platform, two ways to ship.
Hosted on Xiaoxi servers and used via remote API calls — zero ops, elastic scaling, 1000+ concurrent requests.
Deployed on your own servers — on-prem or your cloud — so data never leaves your network.
Try it live for free, or request a trial to connect it to your systems.