Purpose-built models
Not a generic LLM. PhEye serves NLP models trained specifically for PII and PHI entity recognition: higher precision, faster inference, and a tiny fraction of the compute cost of an LLM at the same task.
Why PhEye
Domain lenses
Swap models per workload: general-purpose, healthcare, multilingual, and more. Browse the lens catalog →
Runs in your VPC
Deploy PhEye alongside the data. Sensitive text never leaves your infrastructure: no third-party API, no model-provider account, no outbound dependency.
CPU and GPU support
Standard Docker images run inference on CPU with no special hardware required. GPU-accelerated images (built on PyTorch with CUDA) are available for workloads that need faster throughput or handle high request volume.
Pluggable into Phileas and Philter
PhEye is the default model server for both Phileas and Philter; wire it in via configuration. Or call its HTTP API directly from anything that speaks JSON.
Confidence-aware
Every detection comes with a numeric confidence score between 0 and 100. Tune precision and recall by filtering at a threshold: accept everything above 75, drop everything below 50, decide policy by entity type.
How lenses work with PhEye and Philter
PII lenses are swappable AI / NLP models that plug into PhEye at configuration time. Philter (or Phileas, or any HTTP client) calls PhEye's /find endpoint with text; PhEye runs the loaded lenses, merges their detections, and returns entities with confidence scores. The calling code never has to know which lenses are loaded.
Philter / Phileas
HTTP call with text
PhEye
model server · loads one or more lenses
- General Purpose lens PER · LOC · ORG · DATE · phone · email · URL
- Healthcare / Hospital lens Hospital names · clinical providers · room numbers · medications · symptoms
- Multilingual lenses Spanish PII · French PII · German PII · Portuguese PII · (combinable per workload)
Entities + confidence
JSON response
A Selection of Currently Available Models
Each PhEye Docker image ships with one model baked in at build time. Select the model that matches your language and entity type, or run multiple containers in parallel for broader coverage. Custom models can be developed on request; contact us to discuss your requirements.
| Model | Source | Language | Entities detected | Notes |
|---|---|---|---|---|
| hospitals | Community | English | Hospital names, room numbers, clinical providers | Specialized for healthcare facility identifiers. Detects hospital names, room and ward numbers, and clinical provider references in clinical notes and administrative text. Powered by knowledgator/gliner-pii-base-v1.0 (GLiNER). |
| medical_conditions | Community | English | Disease, disorder | Identifies medical conditions and disease/disorder mentions in clinical or biomedical text. Powered by blaze999/Medical-NER (Transformers token-classification). |
| french_persons | Community | French | Person | Person-name detection in French-language text, using a multilingual news-trained model. Powered by EmergentMethods/gliner_medium_news-v2.1 (GLiNER). |
| french_medical | Community | French | Disease (Maladie) | Medical condition detection in French-language clinical text. Powered by almanach/camembert-bio-gliner-v0.1 (GLiNER, CamemBERT backbone). |
Need a model that isn't listed? Talk to the team about custom model support.
Frequently asked questions
If something here isn’t covered, get in touch and we’ll answer.
What is PhEye?
PhEye is the service that hosts the AI and NLP models that find PII and PHI in unstructured text. POST text to its
/find endpoint and it returns the detected entities with confidence scores. It's designed to plug directly into Phileas and Philter, or to call from any application that needs entity detection over HTTP.How is PhEye different from using an LLM?
PhEye serves NLP models trained specifically for PII and PHI entity recognition, not a general-purpose LLM. For this task that means higher precision, faster inference, and a tiny fraction of the compute cost, and it runs entirely in your own environment with no model-provider account.
What is a lens?
A lens is a swappable model for a particular kind of text: general-purpose, healthcare, or one of several languages. Each PhEye Docker image bakes in one lens at build time, so you run the image that matches your language and entity types, or run several containers in parallel for broader coverage. Browse the options in the lens catalog.
Does PhEye send my text anywhere?
No. Deploy PhEye alongside your data and sensitive text never leaves your infrastructure: no third-party API, no model-provider account, and no outbound dependency.
Do I need a GPU?
No. The standard Docker images run inference on CPU with no special hardware required. GPU-accelerated images, built on PyTorch with CUDA, are available for workloads that need higher throughput or handle high request volume.
How does PhEye relate to Phileas and Philter?
Ready to use PhEye?
Three ways to get going: deploy the open source yourself, spin it up from a cloud marketplace, or work with our team directly. Pick the path that fits.