An engineer builds fein into your stack. It is live in 14 days.

The software is free and you are welcome to run it. This page is the other option. One of our engineers does the build, on your infrastructure, and stays on it afterwards. Your part is two short calls.

Not a support ticket. An engineer inside your stack.

Forward deployed means the person who writes the code sits with the firm that uses it. You get a named engineer, not a queue. They learn how your firm actually works, and they build fein to match it rather than handing you a general-purpose tool and a manual.

It runs on your servers

The engineer deploys into your environment, with your credentials, behind your network. No data moves to us at any point, and there is nothing of ours left running when they finish.

Your ontology, not a default one

Every firm names things differently. Deal stages, ownership, what counts as a pass. The graph is built to your vocabulary, so answers come back in the words your team already uses.

The same engineer afterwards

The person who built it is the person who keeps it running. When a connector breaks or your team starts asking a new kind of question, they already know your setup.

Fourteen days, and two calls from you.

The date is a commitment rather than an estimate. Everything between the two calls is our work, not yours.

Day 1. We connect your systems

On the first call we connect your email, calendar, Drive, LinkedIn and CRM. Around forty minutes, with whoever holds the admin credentials in the room. This runs on your own servers from the first minute.

Days 2 to 13. It reads your full history

fein reads years of email, meetings and documents. Every copy of each person joins into one record. This is the part that takes the time, and it is the part firms doing it alone underestimate: a decade of mail contains a great many people who look like four people.

Day 14. You check it, then it goes live

On a second short call you ask fein things you already know the answer to, and check it against the record. Then it goes live for the team. It holds your full history, like a colleague of many years.

We set it up. You own what we build.

This is the part worth being precise about, because the two halves get confused. What you own is permanent. What you rent is the upkeep.

Yours, whatever happens next

  • The software. Apache 2.0 licensed, on your hardware, under a licence that cannot be revoked.
  • The graph. In your own database. Everything we built during the fortnight stays exactly where it is.
  • No kill switch. Nothing phones home. There is no control plane of ours between your team and their own memory.

What the monthly fee covers

  • Keeping thirty connectors alive. Gmail, LinkedIn, Attio and the rest change their APIs on their own schedule. This is the recurring work, and it is most of what the fee pays for.
  • New answers on request. When the team starts asking something fein does not answer yet, your engineer builds it.
  • Stopping is allowed. Month to month, no lock-in. Cancel and fein keeps running; what you give up is the maintenance, not the system.

$5,000 for the build, then priced on fund size.

Never per seat. Every person at the firm can query the graph, and so can every agent they run.

Fund size
Under $250M$1,000 / month
$250M to $1B$2,500 / month
$1B to $5B$4,000 / month
Over $5BTalk to us
Clone it and run it yourself$0 / forever

An annual plan, paid in advance, costs 15% less. You have thirty days to decide: if fein is not useful to your team we refund everything, the build fee included. The full pricing page

What teams ask before the first call.

Why would we pay for software that is free?

You are not paying for the software. You are paying for fourteen days instead of three months, and for not spending your own engineer on connector plumbing.

fein reads about thirty systems, each with its own authentication, rate limits and idea of what a person is. Getting that running against real data, then reading years of history through it without creating four copies of every contact, is genuinely hard and genuinely boring. We have done it before, which is the entire reason we can put a date on it.

Does our data ever reach you?

No. fein is deployed into your environment and runs there. The engineer works inside your infrastructure with credentials you issue and can revoke. There is no telemetry, no analytics and no control plane of ours. If you disconnect the firm from the internet tomorrow, fein carries on answering.

What do you actually need from us?

Somewhere to run it, admin consent on each source you want connected, and about ninety minutes of a partner's time across two calls. That is the whole ask. Everything else between day 1 and day 14 is our side.

Can we start self-hosted and move to managed later?

Yes, and a fair number of firms do. Run it yourself, see how far you get, and call us when the maintenance stops being a sensible use of your engineer. We pick up whatever you have already built rather than starting again, and the build fee is the same.

What if fourteen days slips?

Then it is our problem, not your invoice. The build fee is fixed and the thirty-day refund covers the whole engagement. We commit to the date because we have run it enough times to know what it takes.

Twenty minutes, and we will walk your stack with you.

We will tell you honestly whether this is a fortnight of our time or an afternoon of yours. If it is the second one, the setup guide is free and public.