FAQ
Tuesday, April 21, 2026
Frequently asked

Questions we keep being asked.

The short version of everything this site tries to answer more slowly elsewhere. If the answer you need isn't here, write to us.

What does Maasv actually do?

Maasv connects to the systems your team already uses, reads them continuously, and builds a single memory of what the business knows.

That memory does three things: it briefs you every morning on what matters, answers questions across every system at once, and closes the follow-ups and obligations you authorize. Nothing happens without you in the loop.

What is context engineering, and is Maasv one?

Context engineering is the category that has formed around getting the right information to an AI model at the moment it matters. The buyer-side vocabulary has settled on four pillars: connected access across the systems an organization runs on, a knowledge layer that resolves entities and relationships, precision retrieval filtered by intent and role, and runtime governance that enforces who-can-see-what at retrieval time and at response time.

Yes. Maasv is a context engineering platform that ships all four pillars. On top of them it runs a cognition loop, which means the brief is the synthesis rather than a list of links, the follow-ups close themselves with your approval rather than waiting for a human to remember, and the memory learns from your corrections so the next answer fits better than the last one did.

Which systems does Maasv connect to?

Email: Gmail, Outlook, Fastmail, HEY, ProtonMail. Calendar: Google Calendar, Outlook, Apple Calendar, Cal.com. Messaging: Slack, Microsoft Teams, Discord. Meetings and transcripts: Zoom, Google Meet, Webex, Otter, Fathom, Fireflies, Gong, Chorus, Grain. Shared drives: Google Drive, OneDrive, Dropbox, Box, SharePoint, iCloud.

CRM: Salesforce, HubSpot, Pipedrive, Zoho, Copper, Attio, Close. Project systems: Asana, Notion, Linear, Jira, Trello, Monday, ClickUp, Airtable, Basecamp, Shortcut. Docs and wikis: Notion, Confluence, GitBook, Coda. Ticketing and support: Zendesk, Intercom, Freshdesk, HelpScout, Front.

Finance: QuickBooks, Xero, FreshBooks, Stripe, Plaid, Mercury, Ramp, Brex, Bill.com, Expensify. HR: BambooHR, Gusto, Rippling, Workday, Deel, Lattice. Dev: GitHub, GitLab, Bitbucket. Marketing: Mailchimp, Marketo, Pardot, Braze. Forms and eSignature: Typeform, Google Forms, Qualtrics, DocuSign, Adobe Sign, Dropbox Sign. Analytics: Mixpanel, Amplitude, PostHog, Segment, Heap.

If you don't see yours, we can add it for your deployment.

OAuth where the source supports it. API and webhook where it doesn't. CSV for one-time imports. No middleware.

Where does my data live?

It depends on the deployment mode. Self-hosted runs the database, the embeddings, the knowledge graph, and the audit trail on your infrastructure, and nothing Maasv stores leaves your perimeter. Maasv-hosted runs the same components on Maasv infrastructure, dedicated to your tenant alone, with secrets isolated on a separate trust tier from the application database. Single tenant, never pooled, in either mode.

Which AI model does Maasv use?

Your choice. Anthropic, OpenAI, or a local model running inside your own network.

Pick a local model and Maasv runs end-to-end inside your environment. Pick a cloud provider and only the specific calls you authorize cross the network boundary, under that provider's terms. Nothing else routes anywhere else. Whichever provider you pick, the model only sees what the role of the person on the other end of the brief is allowed to see.

How are my teams kept separate?

Every team gets its own isolated view, enforced at the data layer before information reaches the agent. Access runs through API keys scoped to tenant and role, or your reverse proxy's JWT if you front Maasv with Cloudflare Access, Okta, or similar. Native OIDC and SCIM are on the enterprise roadmap. Nothing a given person shouldn't see makes it into their brief, their answers, or their actions.

How long does setup take?

Connect your systems on a Monday. By the end of the first full pass, Maasv has already surfaced the things your company knew without knowing: stuck invoices, unspoken commitments, accounts at risk. The specifics belong to you. The shape comes back the same every time.

What size company is Maasv for?

Companies of all sizes. The typical range is three to three thousand employees. The product is built to be worthwhile on day one at the smaller end and still coherent at the larger end.

What if we want to leave?

It depends on the deployment mode. In self-hosted, you already host everything, and you export or erase whenever you want, under your own access. In Maasv-hosted, your tenant data is exportable and erasable on demand, with no waiting period and no negotiation. No lock-in by design in either mode.

How much does it cost?

Pricing depends on the number of connected systems, the scale of your team, and the deployment posture you pick. Write to us with the rough shape of your organization and we'll come back with a straight answer.

How is Maasv different from enterprise search?

Enterprise search helps you find the document that already contains your answer. It covers one pillar of context engineering, precision retrieval over documents, and not the other three. Maasv builds cognition on top of all four pillars, so the brief contains conclusions across the whole picture, not just links inside one of the systems.

Maasv is also dedicated to a single organization, not a shared multi-tenant service. Your cognition doesn't live in a vendor's pool.

Where does Maasv stand on zero-copy federation?

Zero-copy federation has real merits. Querying data where it lives keeps the source-side ACLs intact and means the agent always reads the freshest version, which is the right architecture for some workloads.

Maasv made a different bet at the cognition layer. Connected systems are read once, and their signals are extracted, classified, and connected into a knowledge graph that lives on your infrastructure. The agent then queries that derived layer at memory speed, with role-aware governance enforced on every read.

The honest trade is that the ACL handoff happens at ingest and again at retrieval rather than living entirely on the source side. In exchange, the cognition loop runs over derived structure that the source systems on their own could never produce, and the answer arrives in tens of milliseconds instead of waiting for a federated query to fan out across every source.

Does Maasv replace our existing tools?

No. Maasv reads them. Your team keeps working in Gmail, Slack, Salesforce, and the rest exactly as they do today. The goal is to make the tools you've already bought add up to something more than the sum of their parts.

How do we start?

Write to us at hello@maasv.ai. We'll figure out whether Maasv is the right fit for your organization, and if it is, what the first two weeks look like.