DLRS — Personal OS Documentation
Complete technical documentation for DLRS Personal OS - an autonomous AI-powered personal operating system. Capture via screenshot, email, or voice — AI extracts tasks, links Drive assets, and plans your day.
Who is this for?
What is DLRS Personal OS?
DLRS Personal OS (Personal Operating System) is an AI-assisted productivity platform that acts as yourdigital chief of staff. Unlike traditional task managers where you manually create, organize, and maintain everything, DLRS inverts the model: you provide raw information, the system does the organizing.
Traditional Tools (Notion, Todoist, ClickUp)
- You create projects, tasks, deadlines manually
- You attach files, links, emails one by one
- You set reminders, recurring tasks yourself
- Context switching between apps
- Information lives in silos
DLRS Personal OS
- Paste a screenshot → AI extracts client, project, 3 tasks, deadline
- Forward an email → System creates tasks, finds Drive assets, schedules reminders
- Upload a bank statement → Auto-categorizes transactions, detects subscriptions
- Ask "What should I do now?" → Gets prioritized recommendation with context
- Everything connected in one unified graph
Core Philosophy
DLRS is built on five principles:
1. Capture Everything
Zero-friction input: text, screenshot, PDF, voice, email, URL, file upload. One universal Inbox — no deciding "where does this go?"
2. Understand Automatically
Specialized AI agents classify content, extract entities (client → project → task → deadline → people → priority), and detect relationships.
3. Connect Everything
Tasks, emails, files, clients, projects, calendar events, and financial records form a single relationship graph. No more silos.
4. Act Proactively
Morning briefings, at-risk alerts, subscription renewal warnings, follow-up reminders — the system surfaces what needs attention before you ask.
5. Keep Humans in Control
AI recommends and organizes; sensitive actions (send email, delete file, modify finances) require explicit approval. Full audit trail for every action.
How It Works — The Three-Layer Flow
Every piece of information flows through three stages:
Universal Inbox — Single Entry Point
Key point: You never "create a task." You dump information. The system figures out the rest.
Key Concepts for Beginners
Beginner Glossary
Context Pack — When you open a task, the system auto-assembles everything related: brief, assets, emails, files, checklist. You don't search; it's just there.
Agent — A specialized AI worker with specific tools. Inbox Agent handles capture; Finance Agent handles bank statements; etc.
Orchestrator — The "manager" agent that decides which agent handles your request and coordinates multi-agent workflows.
Provenance/Source — Every fact knows where it came from. "Deadline: Friday" → Source: "Gmail ID abc123, confidence 94%".
Tenancy — Database-level isolation. Workspace A can never see Workspace B's data. Enforced at query level.
The 8 Core Modules
These aren't separate apps — they're views over one unified graph:
🧠 Personal Intelligence
Assistant, Memory, Search, Context Engine, Recommendations. Ask 'What should I do now?' → scored recommendation.
💼 Work Management
Tasks, Projects, Clients, Dependencies, Deadlines, Work sessions, Milestones, At-risk calculation.
📥 Inbox & Capture
Text, Screenshot, File, PDF, URL, Voice, AI extraction, Relationship detection — all in one place.
📧 Communication
Gmail ingestion, Reply generation, Follow-ups, People, Attachments — classified, not just dumped.
📚 Knowledge & Files
Drive indexing, Documents, Notes, Context Packs, Personal search across all sources.
📅 Calendar & Planning
Events, Meetings, Time planning, Daily briefing, Conflict detection, Commitment graph.
💰 Finance
Accounts, Transactions, Expenses, Subscriptions, Invoices, Vault-encrypted statements.
📖 Daily Journals & AI Reflection
Daily synthesis, streak tracking, interactive calendar view, and Tiptap rich editing.
⚙️ Automation & System
Agents, Automations, Jobs, Reminders, Notifications, Integrations, Audit log.
Tech Stack Overview
- Next.js 16.2 (App Router, React 19, Turbopack)
- Tailwind CSS v4 + shadcn/ui base-mira
- Fumadocs UI (Docs layout, Glass theme components)
- PWA (Service Worker, installable, offline-ready)
- next-themes (Dark/light mode, system sync)
Getting Started Path
Follow this sequence for the smoothest onboarding:
1. Installation
Prerequisites, environment setup, database, Redis, and running locally.
2. Quick Start
Create workspace, capture first item, see AI extraction, open Context Pack.
3. Architecture
Understand the data flow, agent system, and entity graph.
4. Deep Dive: Inbox
Master capture modes, extraction review, and automation.
5. Deep Dive: Tasks
Statuses, priority scoring, Context Packs, work sessions.
6. Connect Integrations
Gmail, Drive, Calendar OAuth setup and what each unlocks.
Need help?
- API Reference — All endpoints with examples
- Security & Database — Tenancy, vault, audit
- Changelog — What's new in each version