OpenMed is local-first healthcare AI software that extracts clinical information and removes personally identifying details from clinical text on hardware controlled by the user. Healthcare developers use its Python runtime, Apple Silicon and mobile SDKs, and browser support for on-device clinical NER and PII de-identification.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add maziyarpanahi/openmed --skill running-zeroshot-nergit clone --depth 1 https://github.com/maziyarpanahi/openmedWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/maziyarpanahi/openmed/running-zeroshot-ner)<a href="https://agentmods.dev/skills/maziyarpanahi/openmed/running-zeroshot-ner"><img src="https://agentmods.dev/badge/skills/maziyarpanahi/openmed/running-zeroshot-ner/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/maziyarpanahi/openmed/running-zeroshot-ner"><img src="https://agentmods.dev/badge/skills/maziyarpanahi/openmed/running-zeroshot-ner.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00133 | $0.01759 |
| Opus 5 | $0.00067 | $0.00879 |
| Sonnet 5 | $0.00027 | $0.00352 |
| Haiku 4.5 | $0.00013 | $0.00176 |
Grade A, and why
running-zeroshot-ner scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 158 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Running Zero-Shot NER
Zero-shot NER lets you extract entity types you name at inference time — no
training, no labelled data. OpenMed wraps GLiNER (v1) and GLiNER2 behind a small
index + inference layer, exposed via the openmed zero CLI and the openmed.ner
Python API. It runs on-device.
When to use
- Your label set is custom or evolving ("Device", "Implant", "Allergen") and no fine-tuned OpenMed model emits exactly those labels.
- You have no labelled data to fine-tune with.
- You need a quick prototype or a one-off extraction over an unusual schema.
When to prefer a fine-tuned model instead (extracting-clinical-entities):
for a fixed, well-supported schema (diseases, drugs, anatomy), a fine-tuned
OpenMed model is more accurate and faster than zero-shot. Zero-shot trades some
accuracy for total label flexibility — use it for coverage of new types, then
graduate to a fine-tuned model once the schema stabilises.
Install
pip install "openmed[gliner]" # pulls GLiNER (and GLiNER2 if a recent gliner is installed)
openmed zero deps # diagnostic: prints "GLiNER v1: ok" / "GLiNER v2: ok"
openmed zero deps only checks availability — it does not install anything.
The two-step workflow: index, then infer
GLiNER checkpoints live as local model directories. OpenMed resolves them by a
short model_id via an index.json, so you build the index once and run inference
many times.
openmed zero index <models_dir>— scan a directory of downloaded GLiNER / GLiNER2 checkpoints and writeindex.json(model ids, family, domains, paths).openmed zero infer "<text>" --model-id <id>— run extraction against a model from the index, with labels you supply.
# 1) Build the index over your local models (writes <models_dir>/index.json)
openmed zero index /models/gliner --output /models/gliner/index.json
# 2) Run zero-shot NER with your OWN labels (comma-separated)
openmed zero infer "Patient on insulin glargine via an insulin pump for type 1 diabetes." \
--model-id gliner-biomedical \
--labels "Drug,Device,Disease" \
--threshold 0.5 \
--index-path /models/gliner/index.json
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 158 lines · 133 tokens per session scan A f8ee59e6762c
running-zeroshot-ner is a skill published in the GitHub repository maziyarpanahi/openmed (5,302 stars, last pushed today), licensed Apache-2.0. It adds 133 tokens to every session and 1,759 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
saelens
Train sparse autoencoders to interpret model features.
add-new-model
Use this skill when the user wants to add or port a new model architecture to MLX-VLM — mapping a Hugging Face modeltype to a new file under mlxvlm/models, writing the ModelConfig, matching layer/weight names, reusing a similar existing model, adding a test class, and validating the port. Covers vision-language…
benchmarking
Use this skill when the user wants to benchmark an MLX-VLM change and present the numbers in a PR — fork-vs-main A/B comparisons, isolated-module micro-benchmarks, median-of-N timing with warmup, peak-memory reporting, correctness checks, parameter sweeps, and self-contained reproducible bench scripts to paste into a…
cli-inference
Use this skill when the user wants to run or debug MLX-VLM inference from the command line, including uv run mlxvlm.generate, image/audio/video inputs, local model paths, Hugging Face model IDs, deterministic repro commands, and CLI errors around processors, prompts, model loading, or missing weights.
convert-quantize
Use this skill when the user wants to convert a Hugging Face model to MLX or quantize/dequantize one with mlxvlm.convert, including bits and group size, quant modes (affine, mxfp4, nvfp4, mxfp8), RTN vs AWQ, mixed-bit recipes, dtype casts, calibration (text or multimodal), local vs Hub paths, revisions, uploading to…
server-inference
Use this skill when the user wants to run or debug MLX-VLM server inference, including uv run mlxvlm.server, /v1/models, /v1/chat/completions, /v1/responses, streaming, OpenAI-compatible clients, health checks, metrics, model unload/reload, adapters, trust-remote-code, and server request/response failures.