Launching today
Murph
OpenClaw for your health
4 followers
OpenClaw for your health
4 followers
Murph is a personal health AI that lives in iMessage. It referees health challenges in a group chat with your friends or e-mails a newsletter of your family's weekly health wins. It syncs any wearable, reads all bloodwork, runs health experiments on you to find what works for your body, and helps you build habits that actually stick. It can call the dentist, book DEXA scans, reorder supplements, and basically do anything you want for health. Open source, private by default.
We were using GPT 5.5 for Murph before, and it was too expensive. Now we run most Murph agents through 5.6 Terra, where we're more than comfortable with the balance of speed, cost, and health intelligence. Murph uses Codex CLI under the hood, and 5.6 is great at using subagents proactively. It lets Murph answer users quickly while doing research in the background, which is great for UX.
With 5.6 Sol, we’ve landed 175+ PRs in main and published 81 product changes: Sol/Terra/Luna and reasoning controls; one-time patient-portal imports and biomarker history; Apple Health/WHOOP ingestion and automatic meal-capture recovery; private and group consultation, newsletters, challenges, exact replies, and reactions; and public evidence search for supplements and branded foods. We could finally trust Codex to implement large feature sets and reliably test them end to end. Truly a world of difference both for building Murph and for using 5.6 under the hood. Keep it up!
Hey Product Hunt, rocketman here.
I built Murph because getting healthy is way too hard. Everyone knows roughly what they want to do, but following through sucks. Your data lives in six places, plans don't survive a busy week, and nobody hypes you up. It's lonely. Fixing your sleep or your back pain becomes a second job. Murph's whole pitch is making it easier.
Murph is a personal health AI you text in iMessage, Telegram, or email. It reads your wearables and bloodwork, remembers your context, and reaches out on its own at the right moments. I've been using it myself for months: it ran sleep experiments with me that actually improved my deep sleep, and I do push-ups every day now because if I skip, I get roasted in a group chat with my girlfriend.
What people are doing with it today:
- running health experiments (magnesium for deep sleep, sauna, caffeine cutoffs) to learn what works for their body
- managing chronic pain and tracking what triggers flareups over iMessage
- logging meals with voice memos and photos to watch protein and calories
- finding workout insights buried in mountains of lab and wearable data
- building a physical therapy habit for lower back pain
- having Murph call the eye doctor, book a DEXA scan, and order supplements
- group challenges and a weekly family newsletter (my parents love it)
Murph has its own phone and its own browser. If the job needs a call, it calls. Send it any file (lab PDFs, screenshots, voice memos, a zip of old records) and it makes sense of it.
It's open source (Apache 2.0), self-hostable with one command, and there's a companion app on the App Store. Under the hood: codex cli is the agent runtime and it runs in it's own container, so it is incredibly capable.
Privacy: private by default, encrypted at rest, never used to train models, nothing shared with friends unless you opt in, export anytime.
Would love to hear from the PH community - where have you struggled with health? What makes it difficult to follow through or get started re: healthy habits?
Excited to launch! I build Murph with @rocketman_21 and use it daily to improve my sleep, my environment, and my health overall — it reads my Oura, runs experiments on me like blue light blocking glasses, and tells me what actually works and what doesn't. Hope it helps you too, feedback very welcome!