An ADHD-informed task manager where you never see your task list.
hide-my-list is a conversational AI agent that manages your tasks for you. You tell it what you need to do, and it handles the rest — breaking tasks down, tracking them, and surfacing the right one when you're ready to work. You never look at a list. You never feel overwhelmed by a wall of undone items.
The core insight: for people with ADHD, seeing a long task list isn't motivating — it's paralyzing. Traditional task managers make this worse. hide-my-list takes a different approach.
- Talk to it — describe what you need to do in natural language
- It labels and stores — work type, urgency, time estimate, energy required (in Notion)
- It picks for you — when you have time, it selects a task based on your current energy and mood
- It breaks things down — vague tasks get personalized, concrete sub-steps
- It celebrates wins — completion triggers immediate positive reinforcement
- You never see the list — that's the whole point
This is a Python + LangGraph application backed by Notion (task storage) and Postgres (conversation state, reminder outbox, scheduled jobs). It runs as a Docker Compose stack.
- App: Python 3.12, LangGraph, APScheduler
- Storage: Postgres (checkpointer, reminder outbox, scheduler) + Notion database
- Messaging: Signal via signal-cli bridge (infra-provided)
- Review Pipeline: GitHub Actions with multi-agent Codex review
See docs/architecture.md for the full system design.
git clone https://github.com/NickBorgersProbably/hide-my-list.git
cd hide-my-list
# Install the repo-managed git hooks for this worktree
bash .githooks/install-hooks.sh
# Create .env from .env.template and fill in the values you need
cp .env.template .env
# Start the stack
docker compose up -d
docker compose logs -f appInstall the repo hooks in every worktree you plan to commit from:
bash .githooks/install-hooks.shcore.hooksPath is stored per worktree, so re-run that after each git worktree add.
pre-commit handles fast staged-file checks, and pre-push reruns the
deterministic CI-equivalent checks for changed scripts, docs, and workflow-related
paths so those failures are caught locally before GitHub is the first place they fail.
Every feature is evaluated against ADHD clinical research:
- Executive function support (Barkley model)
- Emotional regulation (Hallowell-Ratey framework)
- Time perception and time blindness research
- Motivation and reward systems (variable ratio reinforcement)
- Cognitive load management
The CI pipeline includes a psychological research evidence reviewer that validates changes against these frameworks.
- Architecture
- AI Prompts (entry point; per-intent prompts live in
docs/ai-prompts/) - Task Lifecycle
- Notion Schema
- User Interactions
- User Preferences
- Reward System
- ADHD Design Priorities
- Security Architecture
MIT — see LICENSE