Last year I posted my pool automation rig (WaterGuru, Hanna, Shelly, ORP, RS485 to the SWG, the works). Good for current state, useless for history. Every spring I'd hit some chemistry weirdness and have no record of what worked last August. Pool Math is the gold standard for dose math but it's stateless, and a paper log doesn't help when the question is "what did I do for this exact problem in July."
So I built PoolQuant. Maintenance log plus an AI assistant that knows your specific pool's full history. Ask it "should I shock?" or "is this algae starting?" and the answer references your actual readings, equipment, and what you did the last time the same thing happened.
The bit that matters for this sub: there's a token-authenticated HTTP API at /pools/<id>/api/log that takes either text or an image. I have an HA automation POSTing my WaterGuru readings to it daily; the same endpoint also accepts photos, so you can fire one when the salt cell display blinks LOW SALT and let the LLM extract the reading off the screen. Entries land with via=api and the token name as the author, so they're indistinguishable from manual entries when you query.
A few example HA snippets:
# Daily WaterGuru push
- alias: Push pool reading to PoolQuant
trigger: time_pattern: hours: 9, minutes: 0
action:
- service: rest_command.poolquant_log
data:
text: >
pH {{ states('sensor.waterguru_ph') }},
FC {{ states('sensor.waterguru_chlorine') }} ppm,
salt {{ states('sensor.waterguru_salt') }} ppm
# rest_command setup (configuration.yaml)
rest_command:
poolquant_log:
url: "https://poolquant.com/pools/<your-pool-id>/api/log"
method: POST
headers:
Authorization: "Bearer !secret poolquant_token"
payload: '{"text": "{{ text }}"}'
content_type: "application/json"
Free, no ads, no email harvesting. Built it for myself; sharing in case it's useful to anyone here.
Would love feedback. Specifically curious whether the "ask questions about your own pool's history" angle is genuinely useful, or whether Pool Math plus a spreadsheet covers it just as well for most folks.
