MorphoGrid FlowOS · field guide

Nineteen ways to turn scattered files into a working data flow.

FlowOS is the multi-file side of MorphoGrid: reconcile, combine, route, validate and deliver datasets without setting up a database. This guide shows what each option solves, what goes in, how it is configured and what comes out.

Explore the options ↓Browse by industry

19purpose-built workspaces

Localbrowser-first processing

Optionalrecipe cleanup per input

No DBfiles in, files out

01 · Combine & compare

Make several files agree on what happened.

⇄

Reconcile

Match expected and actual records, then isolate differences.

Pain point

Two teams send nearly identical lists, and finding missing or changed rows becomes a brittle VLOOKUP exercise.

Use case

Compare a claims register with insurer confirmations before month-end sign-off.

Input · expected / actual

Claim	Expected	Actual
C-101	1,200	1,200
C-102	850	900

Configure

Optional recipe on either file
Match on Claim
Compare Amount

Sample output

Claim	Status	Delta
C-101	Matched	0
C-102	Changed	+50

Outcome: an inspectable exception list and reconciliation export, while both originals remain untouched.

≋

Reconcile+

Handle less-perfect matching with composite keys and tolerances.

Pain point

Invoice IDs repeat across lines and harmless rounding differences create noisy mismatches.

Use case

Reconcile supplier invoices against the ledger using invoice and line number together.

Input

Invoice	Line	Total
INV-9	1	99.99
INV-9	2	45.00

Configure

Keys: Invoice + Line
Tolerance: 0.05
Choose duplicate handling

Sample output

Key	Result
INV-9 · 1	Within tolerance
INV-9 · 2	Missing right

⌘

Multi-file canvas

Prepare each input, then join or append them visually.

Pain point

Every source needs different cleanup before combining, which a single linear recipe cannot express.

Use case

Clean orders and customer masters independently, then enrich orders with customer region.

Inputs

Customer_ID	Total
001	420

customer id	Region
001	West

Canvas

Trim/rename with input recipes
Connect both nodes
Join on normalized ID

Sample output

Customer_ID	Total	Region
001	420	West

＋

Union+

Stack files even when headers are not perfectly aligned.

Pain point

Monthly exports use different casing, spacing or spelling for the same fields.

Use case

Combine regional sales extracts into one quarterly dataset with source lineage.

Inputs

Customer ID	Net Sales
A12	250

customer_id	NetSales
B07	310

Configure

Map equivalent headers
Fill absent fields with blanks
Add source filename

Sample output

Customer ID	Net Sales	Source
A12	250	Jan.xlsx
B07	310	Feb.csv

↗

Enrich

Add trusted reference attributes to a working dataset.

Pain point

A transaction file contains codes but not the names, owners or categories needed to interpret it.

Use case

Add policyholder and broker details to a loss-run export.

Primary + lookup

Policy	Loss
P-44	8,500

Policy	Broker
P-44	Northstar

Configure

Preview both files
Apply optional recipes
Select key and lookup fields

Sample output

Policy	Loss	Broker
P-44	8,500	Northstar

02 · Design & run

Turn one-off work into a visible, reusable process.

✦

AI flow

Describe an outcome and receive a draft workflow to review.

Pain point

Users know the result they need, but not which nodes should be chained.

Use case

Turn “clean files, join customers, route high-value exceptions” into a starter flow.

Input prompt

Clean orders and customers.
Join on customer_id.
Route totals over 10,000.

Draft plan

Input + recipe nodes
Join node
Route + output nodes

Sample output

Orders ─┐
        ├─ Join → Route → Review.csv
Clients ┘           └──→ Standard.csv

Important: AI creates a draft. You review mappings, rules and filenames before running it.

⌘

Graph

Build a workflow by connecting configurable blocks.

Pain point

A process spanning several inputs and outputs is hard to understand as a flat form.

Use case

Model weekly lead processing as input → clean → enrich → split → output.

Input

leads.csv
owners.xlsx
recipe: Normalize CRM export

Configure

Select a block to edit it
Connect compatible ports
Name each output

Graph

Leads → Clean ─┐
Owners ────────┴→ Enrich → Split
                         ↙      ↘
                  assigned.csv unowned.csv

▣

Library

Keep reusable graphs discoverable and versioned.

Pain point

A useful workflow exists only in one browser session or an ambiguously named download.

Use case

Save approved month-end workflows with descriptions and revision notes.

Save

Name: AP month-end
Tags: finance, monthly
Note: Added tax validation

Manage

Search by name or tag
Inspect versions
Export/import definitions

Library card

AP month-end · v3
6 nodes · updated today
[Open] [Export]

▶

Run graph

Bind current files to a saved graph and inspect each node.

Pain point

A repeatable workflow still needs proof that this month’s inputs ran correctly.

Use case

Run the approved AP graph against the latest invoice and supplier files.

Inputs

Graph: AP month-end v3
Invoices: AP_2026-08.xlsx
Suppliers: vendor_master.csv

Execute

Bind each input file
Run supported graph nodes
Inspect counts and errors

Sample result

Node	Rows	Status
Clean	2,490	Passed
Exceptions	31	Ready

03 · Route & calculate

Express business logic without hiding it in formulas.

↙

Split flow

Divide one file into named outputs using clear conditions.

Pain point

Teams repeatedly filter a master export and copy rows into separate files by hand.

Use case

Split claims into high-value review, open claims and closed claims.

Input

Claim	Status	Amount
C1	Open	15,000
C2	Closed	900

Routes

Amount ≥ 10000 → high_value.xlsx
Status = Open → open.csv
Otherwise → closed.csv

Sample output

File	Rows
high_value.xlsx	1
closed.csv	1

Cleanup first: normalize status and amount with an optional recipe before conditions run.

↗

Route+

Use nested AND, OR and NOT logic with route priority.

Pain point

Real exception logic uses overlapping conditions, which can send a row to the wrong owner.

Use case

Route urgent high-value open cases, excluding known test accounts.

Rule

(Amount > 10000 AND Status = "Open")
AND NOT Account starts with "TEST"

Configure

Build nested groups
Order route priority
Name each output

Sample output

Route	Rows
Urgent review	14
Standard queue	207
Excluded tests	5

Expressions

Create calculated fields from readable formulas.

Pain point

Business calculations are buried in copied formulas and become difficult to audit.

Use case

Convert invoice amounts and label material variances.

Input

Amount	FX	Budget
100	3.67	350

Expressions

Local = Amount * FX
Variance = Local - Budget
Flag = IF(Variance > 0,"Over","OK")

Sample output

Local	Variance	Flag
367	17	Over

⌁

Branches

Separate accepted, exception and incomplete records.

Pain point

Good records, genuine exceptions and incomplete rows need different treatment.

Use case

Branch service tickets into within-SLA, breached and missing-due-date queues.

Input

Ticket	Hours	Due
T1	30	24
T2	8

Branches

Hours ≤ Due → accepted
Hours > Due → breached
Due is blank → incomplete

Sample output

Branch	Ticket
Breached	T1
Incomplete	T2

04 · Govern & test

Know what changed, what is expected and what failed.

✓

Contracts

Check incoming files against an expected structure.

Pain point

A vendor silently renames or removes a column and the process fails later.

Use case

Validate the weekly payroll input before transformations run.

Expected contract

employee_id · required text
pay_date · required date
gross_pay · required number

Candidate

Employee ID
PayDate
Gross Pay
Notes

Compatibility

Field	Result
employee_id	Mapped
gross_pay	Type check
Notes	Extra

⚙

Parameters

Replace hard-coded values with named run inputs.

Pain point

Changing month, region or threshold requires editing several rules.

Use case

Prepare one regional sales configuration for each reporting month.

Template

Month = {{report_month}}
Region = {{region}}
Threshold = {{min_sales}}

Run values

2026-08
West
10,000

Resolved preview

Month = 2026-08
AND Region = West
AND Sales ≥ 10000

Current scope: parameter configurations can be previewed and exported for repeatable setup.

⌁

Lineage

Explain how an output value was derived.

Pain point

A reviewer sees a number but cannot tell which source field and steps produced it.

Use case

Trace a normalized claim amount from raw text to final reporting value.

Source

loss_run.csv · row 48
Amount = " AED 1,250.00 "

Trace

Trim text
Remove currency/comma
Cast to number

Output

Claim_Amount = 1250
Origin: loss_run.csv · row 48

◉

Tests

Run repeatable assertions before trusting an output.

Pain point

A recipe can produce the wrong result when source data changes shape.

Use case

Protect a finance cleanup with row-count, null and total checks.

Assertions

Rows ≥ 1,000
Invoice_ID blanks = 0
Amount sum = 2,450,000 ± 1

Run

Load test file
Apply recipe
Compare expected results

Sample output

Test	Status
Row floor	Pass
No blank IDs	Fail · 3
Total	Pass

05 · Deliver at scale

Package the result so the next person can use and trust it.

▣

Deliver

Bundle transformed data with useful supporting artifacts.

Pain point

A bare output file does not explain the cleanup, warnings or counts behind it.

Use case

Hand an audit-ready claims dataset to a reviewer with profile and workflow context.

Input

claims_raw.xlsx
Recipe: Monthly claims standard
Format: XLSX + CSV

Package

Select formats
Include audit details
Name the delivery

Sample ZIP

claims_clean.xlsx
claims_clean.csv
profile.json
workflow.json
README.txt

▶

Batch

Apply the same preparation to a collection of files.

Pain point

The same cleanup must be repeated for twenty branches, periods or vendors.

Use case

Normalize every branch’s monthly sales export and download results together.

Inputs

DXB_sales.csv
AUH_sales.csv
SHJ_sales.csv
Recipe: Branch sales standard

Batch run

Queue multiple files
Apply chosen recipe per file
Review individual status

Sample output

File	Rows	Status
DXB_clean.csv	812	Ready
AUH_clean.csv	604	Ready
SHJ_clean.csv	0	Review

Start with the ugly file.

Clean each input, make the flow visible, and keep the output understandable.

Add to Chrome ↗Ask a question