β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•—  β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•—  β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•— β–ˆβ–ˆβ•—    β–ˆβ–ˆβ•—
 β–ˆβ–ˆβ•”β•β•β–ˆβ–ˆβ•—β–ˆβ–ˆβ•”β•β•β–ˆβ–ˆβ•—β–ˆβ–ˆβ•”β•β•β–ˆβ–ˆβ•—β–ˆβ–ˆβ•‘    β–ˆβ–ˆβ•‘
 β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•”β•β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•‘β–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ–ˆβ•‘β–ˆβ–ˆβ•‘ β–ˆβ•— β–ˆβ–ˆβ•‘
 β–ˆβ–ˆβ•”β•β•β•β• β–ˆβ–ˆβ•”β•β•β–ˆβ–ˆβ•‘β–ˆβ–ˆβ•”β•β•β–ˆβ–ˆβ•‘β–ˆβ–ˆβ•‘β–ˆβ–ˆβ–ˆβ•—β–ˆβ–ˆβ•‘
 β–ˆβ–ˆβ•‘     β–ˆβ–ˆβ•‘  β–ˆβ–ˆβ•‘β–ˆβ–ˆβ•‘  β–ˆβ–ˆβ•‘β•šβ–ˆβ–ˆβ–ˆβ•”β–ˆβ–ˆβ–ˆβ•”β•
 β•šβ•β•     β•šβ•β•  β•šβ•β•β•šβ•β•  β•šβ•β• β•šβ•β•β•β•šβ•β•β•
                
Open Source Β· Self-Hosted Β· Graph Memory

An AI Agent That Builds a
Mental Model of Your Life

Not just chat. Not just memory.
PAAW connects people, work, and context β€” so it actually understands what you mean over time.

  • β†’ Remembers conversations from weeks ago
  • β†’ Connects context across projects, people, and tasks
  • β†’ Executes tasks predictably β€” no random agent behavior

Why PAAW?

[β—‰]β†’[β—‰]β†’[β—‰]

Mental Model β€” Graph-Based Memory

PAAW uses a knowledge graph approach to memory. Every conversation extracts entities, relationships, and key facts β€” your projects, your people, your deadlines, your preferences β€” all stored in a graph database that grows richer over time.

[You] ──works_at──▢ [Company] ──located_in──▢ [City]
β”‚
└──interested_in──▢ [AI/ML] ◀──related_to── [Project]
[ ● ]

Privacy First

Self-hosted. Your knowledge graph, conversations, and personal data stay on your machine. Bring any LLM β€” Claude, GPT, Groq, or run fully local with Ollama.

0/1 *

Scheduled Jobs

Set up automated tasks that run on schedule β€” morning news briefings, investment monitoring, birthday reminders, and more.

MCP

Tool Integration

Connect to web search, Discord, and any MCP-compatible tool. PAAW can search the web, send messages, and take actions.

~/>

Multi-Channel

Chat via Web UI, CLI, or Discord. Talk to PAAW from your phone while commuting. Same context everywhere.

β†’ βœ“

You Define, PAAW Executes

LLMs are probabilistic. Vague goals lead to unpredictable results and runaway token costs. PAAW narrows the probability space β€” you define skills, jobs, and tools in plain English. Structured automation, not autonomous guesswork.

How the Mental Model Works

Every conversation teaches PAAW something new about you. It extracts three things:

[node]

Entities

People, companies, projects, places, dates β€” anything important mentioned in your conversations.

──▢

Relationships

How things connect β€” you work at X, your wife's name is Y, project Z is due on date W.

{ k:v }

Key Facts

Preferences, habits, important details β€” you prefer bullet points, you track crypto, you're learning Spanish.

PAAW Mental Model - Dark Mode

Knowledge graph visualization β€” dark mode

PAAW Mental Model - Light Mode

Knowledge graph visualization β€” light mode

Never Repeat Yourself

Mention a project and PAAW traverses the graph β€” the deadline, the people involved, your last decision. You never start a prompt from scratch.

Nothing Is Forgotten

Three weeks ago you mentioned switching cloud providers. PAAW still knows. Every conversation, every entity, always retrievable.

Same Brain, Any Channel

Start on your laptop, continue from Discord on your phone. PAAW picks up mid-thought. One graph, every channel.

Connections Compound

Mention a colleague and PAAW already knows their role, shared projects, and what you said about the deadline. The more you use it, the smarter it gets.

Get Started in 3 Steps

1

Clone & Configure

Clone the repository and add your LLM API key.

git clone https://github.com/SivaRamSV/paaw.git
cd paaw
2

One-Command Start

The startup script handles everything β€” Docker, database, search engine.

./start.sh
3

Start Chatting

Open the web UI and introduce yourself. PAAW will remember.

http://localhost:8080

Ready to get started?

PAAW is open source and free to use. Self-host it today.

Get PAAW on GitHub