The export is never ready.
Wrong headers. Surprise delimiters. Dates from three centuries. One column called “Count” that somehow means five things.
Local-first transformation studio
Messy data in.
Useful data out.
Clean spreadsheets, reshape exports, join columns, validate records and build quick insights—all from a visual Chrome workspace that business users can actually enjoy.
Your files stay localNo database requiredFree Ollama AI
◈ MorphoGrid
01 Remove top rows
02 Promote header
03 Filter: Revenue ≥ 1k
04 Join columns
Clean preview1,248 rows ready
CustomerRegionRevenue
Acme LabsNorth$12,450
Sprout CoWest$8,190
Orbit WorksEast$6,740
Nova SupplySouth$4,980
No formula archaeology required.
CSV · XLSX · JSON · XML
Pain → agitate → solve
Because “quick cleanup” should not become a three-hour side quest.
Then the spreadsheet grows teeth.
You copy formulas, lose a row, email version_FINAL_v7, and quietly wonder whether anyone has ever understood VLOOKUP.
Build a recipe. Preview. Export.
Drag in the file, select transformation cards, inspect every change and save the workflow for the next equally chaotic export.
Hover or tap the cards
Powerful enough for data work. Friendly enough for Tuesday.
Each tile flips because the honest explanation is often more useful than the brochure version.
⌁
Visual transformations
Filter, sort, calculate, join, rename, regex, date, text, rank, aggregate and more.
Flip for the human translation →Like a tiny data engineer who never says, “Can we circle back after sprint planning?”
◇
Reusable recipes
Save a pipeline as a named card and run it again on the next file.
Flip for the human translation →Teach the cleanup once. Stop performing the same spreadsheet exorcism every Monday.
↔
Format fidelity
Toggle source and table previews, then export back to XLSX, CSV, TSV, JSON or XML.
Flip for the human translation →Your pipe-delimited file can remain pipe-delimited. Somewhere, a legacy system sighs in relief.
⌁
Nested filter logic
Build AND, OR and NOT groups with rich comparisons and unlimited nesting.
Flip for the human translation →“Customers in Dubai OR Abu Dhabi, but NOT test accounts” now fits in a card—not a prayer.
◒
Instant inference
Profile columns, inspect statistics and turn sensible suggestions into pivots and charts.
Flip for the human translation →Ask the dataset what it is trying to tell you before the meeting tells you first.
⌂
Private AI choices
Use local Ollama for free, or explicitly approve cloud providers and choose fallback order.
Flip for the human translation →Local llama first. Cloud brains only when invited. No surprise field trips for your rows.
Productivity levers
Save time where repetitive data work actually hides.
MorphoGrid is not trying to replace a warehouse or an engineering pipeline. It targets the high-frequency gap between receiving a file and being able to use it—the small cleanup jobs that quietly consume hours across a team.
Remove repeated setup
Turn the sequence you perform every week into a named recipe. The next export starts with your steps already understood instead of another round of tribal knowledge.
Lever: standardize once → reuse many timesShorten feedback loops
See the transformed table, native format and step impact before export. Catch the wrong filter while it is still a click—not after the file has reached three inboxes.
Lever: preview early → reduce reworkLower the skill barrier
Business rules become select-and-configure cards. The person who understands the process can build the workflow without translating every request into formulas or code.
Lever: move decisions closer to the expertTurn data into a decision
Column profiles, pivots and chart suggestions help users move past “the file is clean” toward “here is what changed and where attention is needed.”
Lever: compress data-to-insight timeImport
Open the latest operational export.
Apply recipe
Reuse the known cleanup sequence.
Inspect
Review deltas, warnings and totals.
Infer
Create the useful pivot or view.
Export
Return to the source format or choose another.
Where it earns its keep
The recurring file between two systems.
Think weekly CRM exports, monthly finance packs, supplier files, survey responses, campaign reports, HR rosters, audit extracts and JSON payloads that need a human-friendly pass before the next step.
The best fit is work that is too irregular for a permanent engineering pipeline but too frequent to keep rebuilding manually.
A modest example
25 min × 3 × 48If one recipe saves 25 minutes, runs three times a week and is used for 48 working weeks, that is 60 hours returned per year—before counting avoided corrections.Free local AI setup
Connect MorphoGrid to Ollama.
Ollama serves models locally at http://localhost:11434. Because MorphoGrid runs as a Chrome extension, Ollama must allow the extension origin before model discovery and inference can work.
Why the environment variable?
Browsers enforce origin rules.
Browsers enforce origin rules.
OLLAMA_ORIGINS tells Ollama that Chrome extension requests are allowed. For tighter control after installation, replace the wildcard with MorphoGrid’s exact chrome-extension://EXTENSION_ID origin.These steps follow Ollama’s official environment-variable and browser-origin guidance. Restart Ollama after changing the value.
- Quit Ollama from the Windows taskbar tray.
- Open “Edit environment variables for your account” from Windows Settings or Control Panel.
- Create a user variable named OLLAMA_ORIGINS with this value:
chrome-extension://* - Start Ollama again from the Start menu, then pull a model if needed:
ollama pull llama3.2 - In MorphoGrid, open Settings and click Discover models. Select the model from the dropdown.
- Quit the Ollama menu-bar application.
- Open Terminal and set the allowed extension origin:
launchctl setenv OLLAMA_ORIGINS "chrome-extension://*" - Open Ollama again, then download a model if needed:
ollama pull llama3.2 - Open MorphoGrid Settings and click Discover models.
- Edit the Ollama systemd service override:
sudo systemctl edit ollama.service - Add these lines and save:
[Service]\nEnvironment="OLLAMA_ORIGINS=chrome-extension://*" - Reload and restart Ollama:
sudo systemctl daemon-reload\nsudo systemctl restart ollama - Pull a model, then use Discover models in MorphoGrid:
ollama pull llama3.2
Frequently asked, honestly answered
Before you hand it the weirdest file in the shared drive.
Do my files leave the browser?
Normal import, transformation, preview and export happen locally. Local Ollama requests go to your own Ollama service. Cloud AI is optional and requires permission before limited context is sent to the provider you configured.
Is MorphoGrid a replacement for Excel?
No. It is the useful layer before or between spreadsheets: clean the export, standardize the structure, make a repeatable recipe, then send the result wherever it belongs.
Which formats are supported?
Excel workbooks, CSV, TSV, custom-delimited text, JSON and XML. Native-format previews let you swap between source representation and a transformation-ready table.
Why can’t MorphoGrid discover my Ollama models?
Confirm Ollama is running at localhost:11434, set OLLAMA_ORIGINS for the Chrome extension origin, restart Ollama, and then click Discover models. Also verify that at least one model has been pulled.
Does local AI require the Pro plan?
No. Ollama is the free local AI option. Pro adds configurable cloud providers and a sequence of fallback attempts.
What is sent to cloud AI?
MorphoGrid warns you first and asks permission. The request uses column headers, statistical profiles and up to three sample rows rather than automatically uploading the complete file.
Can I reuse a workflow on next month’s file?
Yes. Save it as a recipe card with a name, category, description and tags. Recipes can be searched, versioned, imported and exported without a database.
What happens when a vendor changes column names?
Recipes can map missing expected columns to available columns. You still see the preview and step diagnostics before exporting, so schema drift is visible rather than silent.
How large can a file be?
The free interface currently presents a 100 MB limit. Actual comfort depends on row count, column width, browser memory and transformation complexity because processing occurs on your device.
Where are recipes and API keys stored?
They are stored in the local Chrome extension profile. Provider credentials are not embedded in exported recipes.
From “what is this file?” to “done.”
Install MorphoGrid, drop in a sample, and build your first reusable recipe in minutes.
XLSXCSVTSVCustom delimiterJSONXML