Why Lorah exists
AI work should not reset every time you open a new tool. Lorah was built for people who already use AI seriously and feel the cost of working across disconnected tools.
The work fragments.
One conversation in ChatGPT. Another in Claude. Research in Perplexity. Implementation in a coding assistant. Notes, documents, screenshots, meetings and decisions somewhere else. The models are powerful; the work still fragments.
Serious work needs continuity.
A workspace where your AI work continues across sessions, tools, models, specialists, documents and decisions, and where you decide what becomes the shared truth.
What Lorah is, and is not
A secure, local-first AI workspace where serious work keeps its memory.
- Persistent AI specialists, each with a role and a memory of its own.
- Several model providers in one workspace, on your own keys.
- A review step: you decide what becomes durable project memory.
- A private desktop workspace on your computer. You come back to a project without rebuilding the context.
- A hosted AI model.
- An autonomous agent platform.
- A generic chatbot wrapper.
- A cloud account. There is no Lorah cloud that holds your workspace.
What Lorah costs
The workspace, and nothing else. Model usage is between you and your provider; it never passes through us.
Free forever. Pro for serious work.
Free to download and free to keep using. No card to install, no email to try. The free version is a real way to use Lorah, with limits.
- Pro is for people who want all their specialists active at once, plus Sync, Remember this, Rooms, Background tasks and MCP.
- That is the only thing Lorah charges for: the workspace itself, the memory system, the review flow, the specialist structure, the local app.
Model usage. Not ours to charge.
You bring your own keys for OpenAI, Anthropic, Gemini, Perplexity or xAI. The provider bills you, at its published rate, on its own invoice.
- Lorah adds no markup and takes no commission. Not a cent of that bill reaches us.
- So there is no financial reason for us to push you toward one model over another.
- A model on your own hardware avoids the cloud charge for that call. Lorah labels it unpriced, since there is no verified rate; the hardware is still yours to run.
Who builds it
Lorah is built by LaGuardAI, whose broader work is AI trust, governance, security and control. The same posture, brought to the individual workspace.
Founder. A cybersecurity executive and AI builder with more than 25 years in security, risk, governance and technology leadership. That background shaped the design: know where data lives, know who can see it, know what became durable truth, and keep the human decision-maker in control.
Co-founder. Designs Lorah and builds it, from the first sketch to the code that ships. Builds with AI models and agents every day.
How to try it
Lorah is in beta: real, installable and already useful, and early enough that feedback matters. Do not move your whole life into it on day one.
- Start with one serious projectCreate or import one workspace and add a few specialists.
- Ask a room, or run a SyncPut a real question to the team, or let Lorah propose what the last conversations settled.
- Review the memory Lorah proposesApprove what is true. Reject what is not. Nothing enters without you.
- See whether the next session starts with less re-explainingThat is the test. Your work should compound. Your tools should stop acting like strangers. Your thinking should continue.