Raspberry Pi Offline Health Assistant
Budget: $250 – $750 USD
I have a Raspberry Pi 5 paired with a HealthyPi 5 HAT on my bench and I want to turn it into a completely offline, voice-driven health companion in the next two weeks. All the hardware is already wired up; I just need solid software engineering to pull the pieces together.
Core workflow I’m after
• Speech-to-text: run Whisper locally to listen for conversational prompts and, at minimum, “Give me a status update.”
• Vitals ingestion: poll HealthyPi’s heart-rate, SpO₂, respiration, temperature and any other available metrics in real time.
• Decision layer: generate simple health advice from those vitals (e.g. flagging abnormal ranges) and package it into a friendly sentence.
• Text-to-speech: feed that sentence to Piper so the Pi replies aloud.
• Dual output: show the same advice on an attached display (HDMI or Pi-TFT) while it’s spoken.
Extra touches that would be nice if time allows—basic small talk so the interaction feels natural, plus a modular command structure so I can bolt on more intents later.
Environment & access
• Raspberry Pi OS 64-bit, Python is preferred but I’m open to C++, Rust or whatever you’re fastest with.
• I’ll grant you remote SSH the moment we agree.
• System must boot into the voice service automatically and keep everything offline after initial package installs.
Deliverables
1. Clean, commented source code and requirements file.
2. Systemd (or similar) service so it autostarts on boot.
3. Read-me with setup steps I can rerun on a fresh SD card.
4. Quick demo video or screen recording proving the conversation flow.
Acceptance criteria
• “Give me a status update” is recognized reliably in a quiet room.
• Spoken and on-screen advice corresponds to live HealthyPi readings.
• All processing functions with Wi-Fi disabled after install.
Timeline & next steps
I need this wrapped up inside two weeks. Let me know your availability, estimated hours and rate, and we can kick off right away. Remote testing is fine, and I’m happy to ship logs or run extra diagnostics on my end if you need them.
Core workflow I’m after
• Speech-to-text: run Whisper locally to listen for conversational prompts and, at minimum, “Give me a status update.”
• Vitals ingestion: poll HealthyPi’s heart-rate, SpO₂, respiration, temperature and any other available metrics in real time.
• Decision layer: generate simple health advice from those vitals (e.g. flagging abnormal ranges) and package it into a friendly sentence.
• Text-to-speech: feed that sentence to Piper so the Pi replies aloud.
• Dual output: show the same advice on an attached display (HDMI or Pi-TFT) while it’s spoken.
Extra touches that would be nice if time allows—basic small talk so the interaction feels natural, plus a modular command structure so I can bolt on more intents later.
Environment & access
• Raspberry Pi OS 64-bit, Python is preferred but I’m open to C++, Rust or whatever you’re fastest with.
• I’ll grant you remote SSH the moment we agree.
• System must boot into the voice service automatically and keep everything offline after initial package installs.
Deliverables
1. Clean, commented source code and requirements file.
2. Systemd (or similar) service so it autostarts on boot.
3. Read-me with setup steps I can rerun on a fresh SD card.
4. Quick demo video or screen recording proving the conversation flow.
Acceptance criteria
• “Give me a status update” is recognized reliably in a quiet room.
• Spoken and on-screen advice corresponds to live HealthyPi readings.
• All processing functions with Wi-Fi disabled after install.
Timeline & next steps
I need this wrapped up inside two weeks. Let me know your availability, estimated hours and rate, and we can kick off right away. Remote testing is fine, and I’m happy to ship logs or run extra diagnostics on my end if you need them.
Related categories:
Machine Learning (ML)
Embedded Systems
Software Engineering
System Administration