The product
The questions we get most. For depth, follow the links.
- What is Lorah?
- A secure, local-first AI workspace. You run a team of persistent AI specialists that share project memory you approve. Your workspace, your keys, your machine. OpenAI, Anthropic, Gemini, Perplexity and xAI, on macOS and Windows. Start here.
- Who is it for?
- Founders, consultants, operators, researchers and AI power users who flip between ChatGPT, Claude, Gemini and Perplexity and lose time re-explaining context.
- How is it different from ChatGPT Projects or Claude Projects?
- Four things. Local-first: your workspace lives on your machine, not in their cloud. Multi-provider: five providers in one workspace, no platform lock. Your own keys, zero markup. Reviewed memory with provenance, against their automatic, silent memory.
- Can I use Claude, ChatGPT, Gemini and Perplexity at the same time?
- Yes. One workspace, five providers. Each specialist is bound to one provider at a time; Rooms let different models work the same question.
- What is a specialist?
- A persistent AI colleague with a role, prompt context and accumulated memory. Long-lived: you name it, give it a personality and a mandate.
- What are Rooms?
- Ask the whole team one question at once and each answers on their own, side by side; press Reach a decision and Lorah writes where the room landed, disagreements first. Or Work together, one at a time, each hearing the last. The Rooms guide.
- How does Lorah memory work?
- Specialists propose it, from syncs, rooms and wrapped-up chats. You approve, edit or reject each item in Review. Every approved item keeps its provenance, and when something changes you supersede it rather than delete it. The Memory guide.
- Is Lorah an autonomous agent platform?
- No. Memory, sync briefs, MCP write-backs and task results all pass through Review. It is not built for unattended, long-running action.
- Why use Lorah next to an assistant I already have?
- For projects that outlast one chat. Roles, materials and approved memory live with the project, and you switch models without setting up again. For a quick question, one assistant is enough.
- Does review make an answer correct?
- No. Approving means you accepted it into project memory; it records your decision, not a fact check. Provenance stays visible and any item can be replaced.
- Can I start without the Compiler?
- Yes: pick a starter workspace or build a team by hand. The Compiler needs an Anthropic or OpenAI key or a qualifying local model. Gemini, Perplexity and xAI keys work for specialists, not for the Compiler.
Data and keys
The short version of the trust page.
- What does local-first mean?
- Your project is stored on your computer. There is no Lorah cloud. The network is used for model requests to the providers or endpoints you configured, for licensing and update checks, and for web access or connected tools you use. Crash reporting is optional and off by default.
- Does Lorah store my workspace data?
- No. Our servers see licensing, five fields, and a daily check-in with a random install id, app version, OS, license type and approximate country, never an IP. No prompts, no completions. Crash reporting is off by default and content-free when on.
- How does Lorah use my API keys?
- Stored securely on your computer, in your operating system's keychain. Each call goes direct to the provider, carrying your key to authenticate it. Your keys are never sent to Lorah's servers. Zero markup.
- Can Lorah connect to Claude Desktop, Cursor or Claude Code?
- Yes, through MCP. Lorah can be an MCP server they read approved context from, and a specialist can use an outside tool. Tool calls follow their permission settings. Proposed memory waits in Review. The MCP guide.
- Does a local model make Lorah offline?
- Chats with that specialist stay on your computer. Sync, decisions and the Compiler run on Lorah's own runtime, which can also be a local model. License checks, updates and web search still use the internet. Picking a local specialist does not make every feature local or offline.
Plans and money
Free forever. Pro is $19 a month or $190 a year, for serious work.
- What does Free include?
- Unlimited specialists with three active at a time, fifty memory items, two workspaces, ten documents in Materials, all five providers on your own keys, Review and Memory. Local-first, on your computer.
- What happens after the fourteen-day trial, or if Pro lapses?
- Lorah drops to Free, at the end of the trial or of the billing period. Nothing is deleted. Chat keeps working with three active specialists; Sync, Remember this, new rooms, background tasks and MCP pause until you upgrade. Memory, materials, past rooms and history stay readable, and adding memory by hand still works.
- Do I need a credit card to start?
- No. A card is required only when you upgrade, and you can cancel any time.
- What is BYOK and why?
- Bring your own keys. You pay OpenAI, Anthropic, Gemini, Perplexity or xAI directly at their published rates. Lorah never relays your prompts and adds zero markup. A model on your own hardware avoids a cloud charge for that call; Lorah labels it unpriced, because there is no verified rate, not free. A hosted compatible endpoint can still bill you.
- Can I use my ChatGPT or Claude subscription?
- No. Lorah uses API keys, billed by usage by the provider. A key takes a minute to create.
- macOS, Windows, Linux?
- macOS 12 or later and Windows 10 or later, 64-bit. No Linux for now.
- Where do I upgrade?
- Inside the app: Settings, License & Billing. Checkout and the billing portal open in your browser and land you back in Lorah.
Still didn't find it? The Quickstart covers install, your first hour and every surface. The trust page is the full data story. Or write to support@lorah.ai.