# WellAlly Health: A Local-First Personal Health Record System Built on Claude Code > WellAlly Health is a MIT-licensed personal health record system that runs entirely on your own machine, stores data as JSON and Markdown files, and uses Claude Code slash commands to add records, check drug interactions, and consult across 16 medical specialty domains. No cloud service, no database server, and no account is required. **huifer/WellAlly-health** — Ally-Health is an intelligent healthcare assistant that harnesses advanced AI technology and medical expertise to transform personal health management. Through natural language interaction, it helps users record symptoms, manage medications, organize medical records, and access multidisciplinary health insights—making healthcare more intelligent, p - Repository: https://github.com/huifer/WellAlly-health - Website: https://www.wellally.tech/ - Stars: 957 · Forks: 122 - Language: Shell - License: MIT - Published: 2026-09-15 · Updated: 2026-09-15 · Language: en - Canonical page: https://hysenlabs.com/projects/huifer-wellally-health ## What WellAlly Health is and who it is for Medical records are scattered: hospital portals hold lab results, discharge summaries exist as photos taken with a phone, prescriptions are paper, and daily measurements live in a fitness app. Commercial health apps collect all of this, but upload it to their servers. WellAlly Health takes the opposite approach. The system runs entirely on your own machine through Claude Code slash commands. It stores data as plain JSON and Markdown files in a directory you control. Nothing leaves your machine, no account is required, and no telemetry runs. The README calls this local-first and privacy-first. The README is explicit about what WellAlly Health is not: it is not affiliated with Anthropic or Claude.ai, it is not a medical device, and all analysis is for personal reference only and must never replace professional medical advice. The intended user is an individual who wants AI assistance with health record organization and analysis while keeping full control of their data. ## File-based storage: JSON and Markdown on your own disk The data model is entirely file-based. There is no embedded database, no SQLite file, no server process to start. Records live in a data/ directory as JSON files and Markdown documents. The README describes the directory layout: ``` my-his/ ├── .claude/ │ ├── commands/ │ └── specialists/ ├── data/ │ ├── profile.json │ ├── radiation-records.json │ ├── allergies.json │ └── interactions/ ``` This structure means backups are a standard file copy. The format is described in the README as open JSON plus Markdown, with no proprietary encoding. You can read any record with a text editor, export to other tools, or delete specific files without affecting the rest of the system. The README describes this as yours forever and easy to back up or export. The trade-off is that there is no built-in sync across devices or shared access with a family member or healthcare provider. ## Setting up and running the first commands The only prerequisite is Claude Code installed and opened in the project directory. No npm install, no pip install, and no environment configuration is needed before running the first command. The README gives a seven-step quick start: ```bash /profile set 175 70 1990-01-01 ``` This sets height in centimeters, weight in kilograms, and birthdate. Then: ```bash /save-report /path/to/image.jpg ``` This passes a photo of a lab report to the AI for OCR and structured data extraction. To record a radiation scan: ```bash /radiation add CT chest ``` To start a multi-disciplinary consultation: ```bash /consult ``` That is the full setup. No programming is required to use the system at this level. The system ships 62 slash commands covering everything from basic profile management to specialty analyses for sleep, nutrition, mental health, and pediatric health. ## Drug interaction checking with five-level severity warnings The drug interaction checking feature is one of the most distinctive components of WellAlly Health. The system maintains a local drug interaction database in the interactions/ directory. The README describes five severity levels modeled on a standard classification: A (no interaction), B (minor), C (moderate), D (major), and X (absolute contraindication). The /interaction command checks for drug-drug interactions, drug-disease interactions, drug-dose interactions, and drug-food interactions. The /polypharmacy command handles scenarios with multiple simultaneous medications. This database is a local file, which means it reflects the state at the time you installed or last updated the repository. It does not update automatically from any clinical database. The README does not document the source or scope of the interaction data, so users should treat the interaction check as a reference aid and consult a pharmacist or physician for clinical decisions. ## Multi-disciplinary consultation across 16 specialty domains The .claude/specialists/ directory contains files defining 16 medical specialty domains and an MDT (Multi-Disciplinary Team) coordinator. When you run /consult, the system runs a consultation across all relevant specialties based on your recorded health data and produces a structured multi-specialty response. Single specialty commands like /sleep, /nutrition, /mental-health, /tcm-constitution, /cognitive, /rheumatology, and /rehabilitation address individual domains. Pediatric care is covered through /child-development, /child-illness, /child-nutrition, and /child-vaccine commands. Women's and men's health modules include /cycle, /pregnancy, /postpartum, and /menopause. All analysis runs through the Claude model accessed via Claude Code. The quality of the consultation depends on the completeness of your recorded health data and the capabilities of the model. The README does not describe how the specialty files are implemented, but they are available as plain text files in the .claude/specialists/ directory and can be read directly. ## Limitations: Claude Code dependency, medical disclaimer, and update cadence WellAlly Health requires Claude Code to function. It is not a standalone application and will not run in any other environment. Users who do not have Claude Code installed or who use a different AI assistant cannot use this system without porting the slash commands. The medical disclaimer in the README is substantive, not perfunctory: the system should never replace professional medical advice, it does not produce diagnoses, and all outputs are for personal reference only. The drug interaction database has an unknown scope and update cadence. The last push to the repository was on 2026-07-16. There are no GitHub releases, so version management is through git. The repository is not archived. The system is designed around Claude Code's command-line tooling, and the README notes this explicitly, including that the project is independent and not affiliated with Anthropic. Changes to how Claude Code handles slash commands or skill files could affect how WellAlly Health operates. ## How WellAlly Health compares to Apple Health and commercial PHR apps Apple Health and similar commercial apps aggregate health data from wearables, HealthKit integrations, and manual entry, sync across devices, and share data with healthcare providers through their platforms. The data lives on Apple servers and local device storage under their terms. WellAlly Health makes the opposite trade-off. The data stays entirely on one machine, in plain files you own. There is no wearable integration, no automatic data import from hospital systems, and no cross-device sync. Adding data requires a Claude Code session. For users who prioritize complete data ownership and are willing to manage records manually, WellAlly Health gives capabilities that commercial apps do not: drug interaction checking across all your medications simultaneously, multi-specialty AI consultation contextualized to your specific record history, and an open file format that any tool can read. ## Conclusion WellAlly Health is a practical fit for individuals who already use Claude Code and want a structured, private way to keep medical records, run drug interaction checks, and consult across specialty domains without uploading data to any external service. It is not appropriate as a medical decision-making tool; the README is explicit that the system is not a medical device and all output is for personal reference only. Before adopting it, confirm that the version of Claude Code you have installed supports the slash commands the system relies on, and review the drug interaction database directly at the interactions/ directory to understand its scope. ## FAQ ### Does WellAlly Health upload health data to any server? No. The README states the system is 100% on-device with no uploads, no telemetry, and no external accounts required. All data is stored as JSON and Markdown files on your local machine in the directory where you run the system. ### What AI model does WellAlly Health use for analysis? WellAlly Health uses Claude Code and the AI image recognition model GLM mcp__4_5v_mcp__analyze_image for document understanding, as stated in the README. The system runs through the Claude Code CLI; no separate model is installed locally. ### What slash command checks drug interactions in WellAlly Health? The /interaction command checks for drug-drug, drug-disease, drug-dose, and drug-food interactions using a local database with five severity levels (A through X). The /polypharmacy command handles multiple simultaneous medications. ## Sources - [huifer/WellAlly-health on GitHub](https://github.com/huifer/WellAlly-health) - [Issues](https://github.com/huifer/WellAlly-health/issues) - [License: MIT](https://github.com/huifer/WellAlly-health/blob/main/LICENSE) - [Project website](https://www.wellally.tech/) - [README](https://github.com/huifer/WellAlly-health/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/huifer-wellally-health