Quickstart

Your first 15 minutes in Lorah.

Set up one key, open a workspace, run a real question through your AI team, and approve the memory worth keeping.

Installing Lorah

  1. Download Lorah from our website.
  2. Open the Bond.dmg from your Downloads folder.
  3. Drag the Lorah icon into your Applications folder.

When you double-click the .dmg, this window opens. Drag the Lorah icon onto Applications.

  1. Download Lorah from our website.
  2. Open Bond.exe from your Downloads folder.
  3. The Lorah installer should run.
System Requirements
  • Beta macOS 12 or later (Apple silicon or Intel)
  • Beta Windows 10 or later (x64 or ARM)

When you first open Lorah, you’ll need one API key from a supported provider: Anthropic, OpenAI, Perplexity, or Gemini.

No invite needed — you’re in.

Lorah is free. No account, no credit card, no waitlist. Once it’s installed and you’ve added an API key from a provider you use — Anthropic, OpenAI, Gemini, or Perplexity — you’re running.

Your first 14 days include full Pro, every feature unlocked. After that Lorah stays free on the Free tier (3 teammates, 50 memory items, 2 workspaces) — nothing is deleted, and you can upgrade anytime.

Ready for Pro?

Upgrade inside the app, or manage your plan at license.ai-bond.dev — your Pro license activates under Settings → License.

Manage plan →

Add your first key

When you first open Lorah, onboarding walks you through this. Pick one provider — OpenAI, Anthropic, Gemini, or Perplexity — paste the key, done. Your key stays on your machine.

Look for a short welcome from Yoram waiting in your Review Queue when you finish.

Lorah onboarding — Add an AI provider (Anthropic, OpenAI, Perplexity, Google Gemini)
Onboarding walks you through providers, project context, and your starting team.

Open a starter workspace

Also part of onboarding. Lorah ships with starter workspaces — Personal Board of Directors, CISO AI Risk Operating Room, and more. Pick one and you land in a room that's already set up with teammates and starter memory. Or Build with Lorah AI to design your own team in a chat.

Lorah onboarding — Pick a workspace to start from (Personal Board of Directors, CISO AI Risk Operating Room, Build with Lorah AI)
Each starter workspace ships with a ready-made team and starter memory. Rename, add, or remove anyone after.

Run your first room

Now you're on your own. Open a Meeting or Council and bring two or more teammates in on a real question — one you're actually working on this week, not a toy example. This is where Lorah stops feeling like an app and starts feeling like a working room.

Prompts to try first:

  • “Debate the strongest and weakest parts of this plan.”
  • “What am I missing?”
  • “Research the facts, then have the critic challenge the conclusion.”
  • “Turn this into a decision memo after the team argues it through.”

Prefer a simple start?

Open any teammate 1:1 and chat. Ask them to draft something in their area, or what they think is missing from the project context. The important part is using real work, not a toy example.

Meetings walkthrough Chat walkthrough

Approve & come back

Something useful will come out of your first room. Open the Review Queue and approve one memory candidate — a fact, a decision, a term. That entry becomes part of the project's shared brain.

Close Lorah. Open it tomorrow. The team remembers what you approved, and the next conversation starts from context — not from scratch.

Review Queue walkthrough Memory walkthrough

Explore the app

Here are Lorah's core surfaces. Click any card for a full explanation and walkthrough video.

Ask Lorah

Your command layer. Ask anything about your project. Lorah answers, cites sources, routes you.

1:30 walkthrough
Chat

One-on-one conversations with each teammate.

1:12 walkthrough
Meetings / Council

Bring several teammates into the same room to debate, critique, and synthesize a real question.

1:30 walkthrough
Memory

The reviewed project memory. Facts, decisions, glossary, provenance, and what the team should carry forward.

1:20 walkthrough
Review Queue

Where Lorah proposes what should become durable memory. You approve, reject, or edit before anything is shared.

2:30 walkthrough
Library

Add documents and sources so teammates can work from the material you trust.

0:58 walkthrough
Cost / Spend / Budget

See what each answer costs. Track spend by model, teammate, and surface. Estimate larger runs and set a monthly workspace budget.

Setup guide
MCP setup

Connect tools such as Claude Code, Cursor, Linear, GitHub, and other MCP-compatible systems. Tool actions go through confirmation.

Setup guide

Things to know

Local-first
Your workspace data lives on your machine. The only outbound calls are the provider calls you configure with your own API keys.
Cost visibility is built in
Every reply carries a cost. A monthly workspace budget warns at 80% and pauses at 100%.
First extraction can take a minute
When Lorah proposes memory from a session, it may take 30–60 seconds. That is normal. Nothing becomes shared memory until you approve it.

Next steps

Once you've found your way around, a few places to go from here.

Tell us
Share feedback
Message Yoram on WhatsApp or email hello@ai-bond.dev. We read everything.
Read
How Lorah memory works
The deeper explainer: review queue, provenance, what changes over time, and what stays local.
Read
Why Lorah exists
The story behind the workspace, BYOK, no inference markup, and the LaGuardAI trust posture.

Help

Stuck? Email support@ai-bond.dev.

Found a bug? Message Yoram on WhatsApp or email hello@ai-bond.dev. We read everything.

Discord and bug tracker coming soon. Lorah is built by a small team.