Entity extraction
Names, equipment and dates pulled from text into fields.
We have no published spaCy build. Our AI product reads service pages, job adverts and careers posts through a Claude-compatible API, and we say plainly where a trained pipeline would help.
Names, equipment and dates pulled from text into fields.
A FastAPI service that runs a trained pipeline.
The sentence that produced a value kept next to it.
Moves onto the 3.8 line.
Illustration. A dashed outline marks where the technology would sit. We have no published build on it.
A maintenance release.
A build-configuration release.
Marked latest on GitHub.
Source: spaCy releases on GitHub, checked 6 November 2026. The release headers show day and month, and the asset timestamps give 2026 for 3.8.18.

The engine crawls a site, reads service pages, job adverts and careers posts, and writes a graded brief with the evidence beside each score.
A Python 3.13 back end with telematics ingest and PostGIS.
Python
PyTorch
Docker
scikit-learn
TensorFlow
Apache AirflowspaCy 3.8.18, marked latest on GitHub and released on 26 October 2026. Version 3.8.17 came out the same day and 3.8.16 on 24 August.
No. We say so on the first call and show the nearest work, which is Sales OS, which reads service pages and job adverts through a model API.
You install spaCy, then download a trained pipeline, for example with python -m spacy download en_core_web_sm.
You do, from the first commit. Repositories and cloud accounts stay in your name.
Thirty minutes on your documents, your fields and your spaCy version. No sales pitch.