jigripokri

jigripokri / POHA

Public

Personal Overnight Helper Agent. Runs while you sleep. Serves up a morning brief before your alarm.

ai-assistant claude claude-code morning-brief personal-assistant
30
7
85% credibility
Found May 20, 2026 at 40 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

POHA (Personal Overnight Helper Agent) is a personal AI assistant that runs scheduled tasks while you sleep and delivers a morning briefing before you wake up. It connects to your existing services—email, calendar, messaging apps—to observe what's happening in your life, maintains persistent memory about you and your priorities, and acts on a schedule you define. The system includes seven pre-built scheduled tasks (morning briefing, evening reflection, weekly review, commitment tracking, task planning, weekend activities, and monthly finances) plus five on-demand skills for drafting messages, generating replies, finding insights, creating task lists, and group-chat humor. All your data stays on your machine; the AI reads your information to serve you better but never shares it. The project is designed for people managing complicated lives who want an AI that works proactively in the background rather than waiting to be asked questions.

How It Works

1
💡 You discover an AI that works while you sleep

Someone tells you about POHA—an assistant that runs overnight and delivers a morning briefing before your alarm goes off.

2
📱 You install the app and open your new assistant

You download Claude Desktop, clone the project folder, and open it. Your new AI chief of staff is ready to meet you.

3
✏️ You tell it about your life and priorities

You fill in a simple form: your name, your partner's name, who matters most to you, and how you want your assistant to sound.

4
🔗 You connect your email, calendar, and messaging apps

You link the services you already use—Gmail, Google Calendar, WhatsApp—so your assistant can see what matters to you.

5
🌙 Your assistant learns your world overnight

While you sleep, your AI reads through your messages, emails, and calendar. It updates its memory and prepares your briefing.

☀️ You wake up to your personalized morning briefing

Your phone chimes with an email: your commitments, unanswered threads, and what's on your plate today—all organized by what matters most to you.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 40 to 30 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is POHA?

POHA is a personal AI agent that runs overnight and delivers a morning briefing before your alarm goes off. It connects to your WhatsApp, Gmail, Calendar, and other services to synthesize everything happening in your life, then emails you a prioritized summary with one-tap action links. The system runs scheduled tasks at specific times (5am briefing, 9pm reflection, weekly reviews) and offers on-demand skills like /draft, /reply, and /gsd for immediate needs. It leverages your existing Claude subscription through a persistent memory system built on plain markdown files.

Why is it gaining traction?

The concept fills a real gap: an AI that works while you sleep and surfaces what matters most before you check your phone. The tier-weighted attention system automatically prioritizes messages from your partner over group chat memes. The acknowledgment loop via email is clever -- you tap a link to mark things done, and POHA learns. For developers already deep in the Claude ecosystem, this bridges the gap between a reactive chatbot and a proactive personal chief of staff.

Who should use this?

Developers managing complex lives with multiple commitments who regularly lose threads across WhatsApp, email, and tasks. It's ideal for those already paying for Claude Pro or Max and wanting to extend it into a persistent second brain. You need to be comfortable writing context in markdown and willing to invest a few hours upfront in setup.

Verdict?

This is a genuinely interesting concept with thoughtful architecture, but 30 stars and a 0.85% credibility score mean it's early and unproven. The README references several documentation files that don't appear to exist in the repo, which raises questions about completeness. Worth watching and experimenting with, but wait for more community validation before committing your morning routine to it.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.