Connect it once. The memory builds itself.

Every document, email and database row is read as it arrives and filed with the customer, order or project it belongs to. Nothing to set up first.

Eight sources that never mention each other become one project.

A change order sits in an email, the site protocol is a scan, the milestone invoice is a PDF and the payment lives in a database. fluctor recognises that they all belong to project P-2210 and files them together, with every source kept.

Two people reviewing a contract

contract_P-2210.pdf

CONTRACT · SHAREPOINT

Container ships at a crane

RE: change order #3

EMAIL · OUTLOOK

Aerial view of a construction site

site_protocol_14-03.jpg

PROTOCOL · SCAN

Warehouse aisle stacked with pallets

Delivery #2210-7

DELIVERY · PDF

Shelf of binders

invoice_2088.pdf

MILESTONE INVOICE · PDF

Laptop showing a financial dashboard

Payment · open

OPEN ITEMS · SQL

Laboratory workspace

inspection_report_9.xlsx

QUALITY · EXCEL

Robot arms on a production line

RFI #14, bracket spacing

EMAIL · GMAIL

Change order #3 was agreed by email on 2 April. The milestone invoice of 4 June does not include it: €8,400 not billed.

One file per customer, kept up to date by itself.

No record to create and no fields to maintain. Everything that mentions a customer, an order or a project lands in its file the moment it arrives, whether it came as a PDF, a scan, an email or a row in your database.

Halden Steel Ltd

customer since 2019 · Sheffield · 2 contacts

Documents134invoices, orders, delivery notes, emails, one framework contract
Open orders3#88, #91, #94 · €61,200 in total
Open invoices2invoice_2041 · invoice_2088 · €50,400 outstanding
Contractframework_2024price review 60 days before anniversary · notice period 3 months
Last contact19 MayRE: order #88, partial delivery
Open signals2delivery without invoice · orders down against own history

Every fact keeps the line it was read from.

A payment term, a signature date, an agreed amount: each one is stored together with the document and the sentence it came from, so anyone can check it in seconds.

Payment terms, Halden Steel Ltd: 30 days net

“…payable within 30 days of the invoice date.”

framework_2024.docx · page 3

Delivery #88-1 signed for on 14 May

“Received in full, 40 units.” with signature

delivery_88-1.jpg

Change order #3 adds €8,400 to project P-2210

“…additional steel bracing as per attached, EUR 8,400 net.”

RE: change order #3 · email of 2 April

Ask how it looked on the first of May.

The memory keeps its history. Pick a date and see an order, a customer or a project as it stood then, not as it stands today. Useful when a customer disputes what was known when, and for a new colleague catching up on a project. How far back it keeps is set per company.

1 Mar
1 Apr
1 May
1 Jun
today

AS OF 1 MAY

Order #88 · status: shipped

Delivery #88-1 · signed 14 May: not yet

Invoice: none on file

TODAY

Order #88 · status: invoiced

Delivery #88-1 · signed 14 May

invoice_2088 · issued 4 June · open

Eight ways in. Then it keeps itself current.

Connect the places where the work already happens. From then on every new file, message and row is read as it lands. No exports, no uploads by hand.

Microsoft 365

SharePoint

Outlook

Microsoft 365

Teams

Microsoft 365

Google Drive

Workspace

Gmail

Workspace

SQL

PostgreSQL, MySQL

Any API

REST

Webhooks

JSON push

Files you already have can also be dropped in directly: PDF, scans, Word, Excel, email files, images.

LIVE · THE MEMORY AS A MAP

/graph

Product screenshot

CUSTOMERS · ORDERS · DOCUMENTS, CONNECTED

The same memory, drawn as a map. Groups that belong together sit together, and a click opens the file behind any point.

This memory is what your people ask, what the signals watch, and what your agents run on.

Twenty documents are enough to see it.

Bring them to the call. Watch them turn into one file per customer, with every source still attached.

No integration project to start  ·  Your data stays in the EU