The questions we'd ask too.

You can, and many teams start there. It costs you twice. Every question pays for everything the agent pulls in, about eight times more per answer than asking Mithrandir on our own demo. And the agent gets the raw mess: scattered threads, duplicate docs, stale files, with nothing organized and nothing checking who may see what. Mithrandir organizes all of it into one structure and hands each person only what they can access at work.

No. Mithrandir is the whole company's memory: sales, legal, finance, people, engineering, and so on. The same memory holds a sales discount, a contract clause, a hiring plan and the code behind a feature, each organized where it belongs and each visible only to the people allowed. One seat price covers the whole company.

Connect the sources where your work already happens: Slack, email, docs, tickets, calls, code. Mithrandir builds the memory from your history, not from a blank page, so the first useful answers come back the same week. Your agents reach it over MCP, so nobody changes the tool they work in.

No, and that's the point. Every knowledge tool that depends on someone filing a doc eventually goes stale, because the filing is the first thing to get dropped. Mithrandir reads the places where work already happens (threads, emails, tickets, docs, meeting transcripts), so nobody writes the same thing down twice, and nobody has to remember to. The answers then come back inside the tools your team already works in.

Then nothing can recover it, and we won't pretend otherwise. But most of what a company knows is written down somewhere; it just lives in a thread, a ticket comment, an email, a transcript, and then gets buried. Mithrandir reads it where it already lives and keeps it organized, so the answer is there the next time someone asks.

The new version becomes the answer and the old one stays on record with its dates. An amended contract supersedes the clause it replaced, and the replaced terms still answer for the period they covered. When someone with access flags a result as wrong, the correction becomes current and the old one is retired: kept for the history, never served as current again. Nothing is silently overwritten, and nothing old comes back dressed as new.

Mithrandir follows each tool's own access rather than inventing new rules, so a person sees in their agent exactly what they can already open at work. Where a tool has no rules, your company's hierarchy decides, and ethical walls you set apply on top of everything. When someone leaves a channel or is removed from a tool, their access follows.

We don't train models on customer data. Mithrandir runs in your own cloud account (AWS, Azure or GCP), set up with our Terraform, and the model runs on your own provider account or on Amazon Bedrock in your own AWS account. We never host your data, so restricted content never has to leave your infrastructure. If your organization has hard requirements beyond that, bring them to the first call.

Thirty minutes with the founders, not a scripted sales demo. Bring one workflow where context keeps getting lost. We show you what each of your people would see in that workflow, talk through which of your sources it would connect to and how it would sit in your infrastructure, and answer your security questions directly. There is no form to fill in first.