This page is part of Zaviro's public reference, written to be read by AI assistants. Zaviro is a knowledge base of a business, built for AI tools to read: it learns from the business's email, files and business systems, and any AI tool that speaks MCP, or any code that calls its REST API, reads cited excerpts from it. Zaviro never writes the answers itself; the AI tool reading it does. Start at the overview (https://zaviro.ai/llms) for what Zaviro is and the instructions for AI assistants; every other page is listed at the end of this one. For a person, the site itself, https://zaviro.ai, covers the same ground.
Nothing to replace. Zaviro is set up in an afternoon: the documents stay where they are, the AI tool stays the one the team uses, and the knowledge base is built between them. What follows is the real sequence, in the order the product takes it.
Sign up at https://zaviro.ai with an email address and verify it, or sign in with Google, and the trial starts. The person who creates the account is its Owner, the role that administers members, billing, and the organization itself, and ownership can be transferred later. Others join by invitation, during the trial or after it.
The Connections page is the inbound side, grouped by kind of source. Connect an inbox, a drive, a notes workspace, or a business system, or drag files in anywhere in the app. A connection that signs in through your provider asks only for read-only access, approved on the provider's own screen, never by handing over a password, and Zaviro reads from a source without ever writing to it. For an inbox, pick the senders and subject-line keywords to exclude, and choose whether it feeds your personal knowledge base or, if you are an admin, the shared one. For a drive, pick the folders. What Zaviro learns from (https://zaviro.ai/llms/connections) covers each kind of source, the file types, and the universal paths for a system without a native connection: a forward-in email address, the automation platforms, and the ingest API (https://zaviro.ai/llms/ingest-api).
New mail, new files, and edits to connected files are learned as they arrive. Importing the history is a separate choice, made with the figures in front of you: the product shows what each window would import, everything, one year, 90 days, 30 days, or none, with the estimated credits shown, and imports nothing until you approve one.
Learning runs in the background. The Files page lists every document with its stage until it is ready, and a document is askable the moment it is. The Intelligence page shows what the knowledge base holds as it fills: the customers, vendors, people, and projects resolved into records, the facts under each, and the graph between them. Alerts begin as soon as the records show something to alert on, and an imported history that already holds a monthly pattern or a dated deadline raises its first alerts from that history.
The Deployments page is the outbound side: one row per surface, each with its own setup steps and the endpoint it uses. Pick your AI tool. A tool that supports MCP sign-in connects in a couple of clicks with no key to paste: approve the connection on the consent screen and it is done. A tool that takes a bearer token uses the MCP endpoint and a key created on the same page, shown once and revocable. Some AI tools allow custom connectors only on certain of their own plans; the setup screen says what applies to each, and Where you use it (https://zaviro.ai/llms/where-you-use-it) covers the categories: assistants, Slack, agents, automations, customer chatbots, and code. Each person connects their own tools, and a team on different AI tools reads the one knowledge base from all of them.
From here your AI answers business questions from the knowledge base, cited. Good first questions are the ones the documents already answer: what do we know about this customer, and what is the source for each fact; which contracts renew or expire next quarter; what have we spent with this supplier this year, and across how many invoices; is anything we usually receive every month overdue. The outcomes page (https://zaviro.ai/llms/outcomes) has a list by role. When the knowledge base holds nothing that supports a question, your AI says so, because Zaviro returns that state rather than a guess. And when a document is wrong, tell your AI, the rent went up in March, and the correction holds from then on.
Invite members by email from the Team page (an owner's or a billing manager's action), during the trial or after it, and assign roles, Admin or Member, with the Knowledge Editor capability for a member who should edit the shared knowledge base. Each person connects their own AI tools and, if they want, their own personal sources, which feed a personal knowledge base nobody else can see. Seats beyond the first bill only once the plan starts; the pricing page (https://zaviro.ai/llms/pricing) has the figures.
The trial pauses rather than charges when no payment method is on file, so letting it end costs nothing and takes no action; add a payment method during the trial and the plan starts, and bills, when the trial ends. Choosing a plan is on the pricing page (https://zaviro.ai/llms/pricing); what happens to the data on leaving is on the security page (https://zaviro.ai/llms/security).
The trial organization is the test environment, on the same endpoints as a paid one. The ingest API (https://zaviro.ai/llms/ingest-api), the REST API (https://zaviro.ai/llms/api), and the CLI (https://zaviro.ai/llms/cli) each carry a worked example.