Setup in Minutes

Local AI Setup Made Easy

Download the installer and Robbie does the awkward setup work: checking your hardware, recommending a suitable local model, setting sensible defaults and launching a private local chat experience.

1
Download the installer
Run the installer on Windows, Mac or Linux. Sensible defaults are pre-selected — change them if you want to, or keep moving.
2
Run the onboarding
AGENT ROBBIE profiles your CPU, RAM and GPU, then recommends a local model that should run smoothly on your machine. Introduce yourself while it installs.
3
Your AI starts itself
The local LLM server, the embeddings server and the workspace API launch as background services. No terminals to keep open — it just works.
4
Connect tools when needed
Start with simple local chat. Add email, calendar, GitHub, web research, Image Studio, documents or Agent Memory when you want Robbie to help with a real workflow.
Everyday Work

Connect the Apps You Actually Use

Robbie is designed to help with normal work: messages, meetings, files, code, web research and project memory. Connectors stay off until you add credentials and approve the action.

Email

Draft, summarise and organise mail

Use Gmail, Outlook or local Thunderbird profiles to review messages, draft replies and prepare follow-ups. Sending still asks for approval.

Live / staged by provider
Calendar

Plan days and prepare meetings

List events, create calendar items, prepare agendas and keep scheduling work visible before anything is changed.

Google Calendar live after OAuth
Files & Memory

Use your local knowledge

Search Agent Memory, local notes and project documents so Robbie can work from your real context instead of guessing.

Local-first
GitHub

Read, update and ship code

Search repositories, read files, write content, open pull requests, merge approved work and trigger GitHub Actions.

Repo automation ready
Web Research

Search current information

Use Firecrawl, Tavily, Brave Search or fallback search to check the web when a local model needs fresh facts.

Provider keys optional
Work Tools

Grow into your workflow

Slack, Obsidian, VS Code, Cloudflare, CRM, analytics and publishing integrations are available as your work needs them.

Permission-gated
What You Get

Private by Default. Personal by Design.

Six things every AGENT ROBBIE install does out of the box — so your AI fits your machine, your work and your way of thinking.

Auto Hardware Check — Curated Install

AGENT ROBBIE profiles your CPU, RAM and GPU during install and recommends a model that will run smoothly on your machine. CPU-only laptop, gaming rig, or workstation with a discrete GPU — you get a setup tuned for your hardware, not a generic default.

Chat First. Tools When Useful.

Out of the box you get a fast, focused local chat. Connect email, calendar, web search, GitHub, notes and other services only when the job needs action. Robbie keeps simple prompts light and loads tools deliberately.

Remembers How You Work

A private memory database lives on your disk — not in someone else’s cloud. AGENT ROBBIE remembers your preferences, your projects, your tone of voice and the way you like to work, so every conversation picks up where the last one left off.

Keeps Your Secrets, Locally

API keys, OAuth tokens and integration credentials live in an encrypted local vault — never in chat history, never on our servers. AGENT ROBBIE uses them at request time without ever exposing them in plain text.

Cloud When You Want It. Not When You Don’t.

Conversations, files and models stay on your disk by default — you can even work fully offline. When you do want a frontier model, bring your own API key and AGENT ROBBIE will route that one request direct to the provider. No middleman, no relay, no surprise bill.

As Technical As You Want To Be

Run AGENT ROBBIE exactly as it ships, or open the hood: swap LLMs, rewrite the onboarding, pick your own integrations, tune system prompts and routing. Friendly defaults for newcomers, full control for power users.

What You Need

Most Modern PCs Can Run Local AI*

Smaller local models can run on a CPU — a GPU just makes them much quicker. AGENT ROBBIE auto-detects your hardware and recommends a model that fits the hardware available. *Please check hardware requirements before purchase if fully local AI is important to your workflow.

Minimum (CPU only)

  • Windows 10/11 64-bit
  • i9 CPU
  • 16 GB RAM
  • 100 GB+ free space
  • Internet for initial install & model download

Recommended (GPU)

  • Windows 11
  • 16–32 GB RAM
  • NVIDIA RTX (8 GB+ VRAM) or AMD RX 7000-series
  • SSD with 100 GB+ free space for the model cache