Dev Tools
VelesDB
Explainable, local-first memory engine designed for AI agents.
★ 97 starsNOASSERTION
View guide →Local-first healthcare AI: clinical NER and HIPAA PII de-identification that runs 100% on-device.
openmed is an open-source dev tools project by maziyarpanahi (maziyarpanahi/openmed on GitHub). It has 5,463 GitHub stars and 700 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 8 October 2026.
Decide whether you need PII de-identification, clinical entity extraction, or both, and on which platform — server, desktop or mobile.
Install the openmed Python package with the extras for your hardware (MLX on Apple Silicon, or the Hugging Face/PyTorch path).
Fetch the model artifacts for your task; after this one-time download everything runs offline.
De-identify a synthetic note and extract medical entities locally, then review the spans and redacted text.
Detects and removes HIPAA-relevant PII from clinical text entirely on-device — no cloud, no patient data leaving your network.
A catalog of 2,200+ medical models covering clinical entity types in 21 languages.
An MLX runtime (with CoreML fallback for supported artifacts) accelerates token classification on Apple Silicon; PyTorch checkpoints can fall back across hardware.
OpenMedKit for Swift/iOS and an ONNX-based Android runtime keep tokenizer offsets and span boundaries aligned with the Python runtime.
Ships portable Agent Skills for de-identification, clinical NER and FHIR export that load into Claude Code, Codex, OpenCode and compatible agents.
The same ONNX models can run in the browser, and optional service extras expose OpenMed as a local service.
Clone OpenMed to explore the source, models and skills.
git clone https://github.com/maziyarpanahi/openmed.git
cd openmedInstall the Python package with the MLX extra on Apple Silicon, or the Hugging Face extra elsewhere.
pip install --upgrade "openmed[mlx]"
# elsewhere: pip install --upgrade "openmed[hf]"Use the documented API to de-identify text and extract entities entirely on-device.
python -c "import openmed; print(openmed.deidentify('Patient John Smith, MRN 12345'))" # see README for the exact commandOpenMed is maintained by maziyarpanahi and contributors, with the project site at openmed.life.
Yes — it is open source under the Apache-2.0 license, including for on-premise use.
No. After a one-time model download, inference runs 100% on-device with no cloud calls and no API keys.
Cloud APIs (such as Comprehend Medical) require sending PHI to a third party; OpenMed keeps all processing inside your network or on the user's device, which simplifies HIPAA posture.
pip install the extra for your hardware, download a model, and run the de-identification example on synthetic text first.
Explainable, local-first memory engine designed for AI agents.
Agentic IDE that runs 100+ parallel coding agents in isolated workspaces on your own subscriptions.
A local-first database client with AI SQL help — 40+ databases, desktop, web, Docker and CLI
Generates agent-first Go CLIs (plus MCP servers) for any API or website — with local SQLite sync built in
A local-first Agent OS — drive Claude Code, Codex and OpenCode from your browser.
Local-first web search, fetch, crawl and research for AI coding agents over MCP — no API keys.
The most-starred other Dev Tools projects in the LibreAI catalogue — the realistic shortlist if openmed is not the right fit.
The open source coding agent
The model-definition framework behind modern AI — thousands of pretrained transformers for text, vision, and audio.
Anthropic's agentic coding tool that lives in your terminal and works across your whole codebase
Cross-platform desktop all-in-one manager for Claude Code, Codex, OpenCode and other coding agents
OpenAI's lightweight coding agent that runs locally in your terminal
Free and open — grab the source, star the repo, and start creating.