Skip to content
health bot for creating habits
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
Habot Data Model.png


Habot is a coach for creating daily healthy habits through a conversational interface.

As you’re onboarded with Habot, you rate how well you’re doing in certain life areas. Habot then recommends a set of habits that would be best to work on. After committing to a habit, you track progress with Habot every day through delightful text message exchanges. You earn a new streak every 3 consecutive successful days and are encouraged to collect at least 7 streaks to create the new habit (21 days in total).

Habits included are built off of BJ Fogg's Tiny Habit principles ( - that is they are:

  • incredibly specific
  • can be done in 30 seconds or less

If you have examples of Tiny Habits you would like added to Habot, please email me!

Current features:


  • Mobile phone number + country.
  • Timezone.
  • Name.
  • Ratings for 'profile' (health categories: mind, sleep, relationships, performance, movement, nutrition).
  • Habit selection.
  • Time to be reminded.


  • 'Passwordless' sign up and sign in via a 6 digit randomly generated pin sent to verify mobile number.


  • Data visualization to track progress toward 3 day streaks and number of 3-day streaks (with encouragement to collect 7 x 3-day streaks or 21 days total).

Conversation features:

  • Daily reminders sent @ specified time in local time zone.
  • Success tracking via reply with '#yes'.
  • Pausing of habit tracking if inactive for > 7 days. Reactivation with 'unpause'
  • Small talk :)

API integrations:

  • Twilio
  • DialogFlow

Tech stack:

  • PostgreSQL
  • SQL Alchemy
  • Python
  • Flask
  • Jinja
  • JQuery
  • Bootstrap
  • CSS
  • HTML
  • JavaScript

Libraries utilized:

  • schedule
  • arrow
  • phonenumbers
  • pycountry
  • pytz
  • emoji
  • D3
  • TypeIt

Next Features:

  • testing framework that forces specific servertime to automate test user cases
  • handling all errors during sign up flow
  • fixes with the scheduling utility / thread
  • preparations for deployment to

Video of current local build:

Data model:

data model

Notes on cloning:

  1. You'll need a Twilio and DialogFlow account.
  2. At a minimum, your DialogFlow agent needs to be programmed to recognize '#yes' and 'unpause'.
  3. To process user responses, configure DialogFlow for fulfillment. I used ngrok to create a webook at '/'.
You can’t perform that action at this time.