Clean and transform CSV data with AI, locally
Describe the transformation you need in plain English. The agent writes and runs the Python script, shell pipeline, or Node job — you just approve and get clean output.
- 1-click uninstall
- Cancel anytime
- Files never leave your computer
Impact
What changes
The same task, two ways — how it plays out by hand today, and what changes once Lapu AI runs it for you.
Without Lapu AI
A data analyst spends 20-40 minutes writing a Python script or fumbling with spreadsheet formulas to clean and transform a CSV. Sensitive data gets uploaded to online tools for quick fixes.
With Lapu AI
Lapu AI generates and runs the transformation script in minutes. The original data is preserved, the output is validated, and processing happens locally on your machine.
The challenge
Processing local data files often requires writing one-off scripts, switching between spreadsheet applications, or uploading sensitive data to online tools. Teams working with customer data, financial records, or proprietary datasets need a way to transform and analyze files without sending them to third-party services.
How Lapu AI solves this
Tell Lapu AI what you want done to your data (drop the empty rows, merge these files, summarize by month) and it does it. No code, no fighting with spreadsheet formulas. It shows you exactly what it is about to run, hands you clean output, and tells you what changed: row counts, anything that looked off. Your data never leaves your machine.
Data processing runs locally via scripts the agent generates. Only structural context (like column names) is sent to the AI model.
Workflow
How it works
Load your data files
Point Lapu AI at your data files. The agent reads the first rows using File Read to understand the schema, column types, and overall structure.
Describe what you need
Tell the agent what to do in plain language — filter, merge, aggregate, reformat. The agent writes the appropriate script (Python, Node, or shell) and shows it to you before execution.
Run, validate, and export
Approve the script and the agent runs it — either through sandboxed Python/Node execution or shell pipelines. It reads the output, reports row counts, shows a sample, and flags anomalies.
Under the hood — for the technically curious
Behind the plain-language request, Lapu AI writes and runs the actual processing: shell pipelines (awk, jq, sed) for quick jobs, or sandboxed Python and Node scripts for heavier transforms, with automatic package management. The sandbox runs locally with output capture, timeouts, and validation. Only structural context like column names is sent to the AI model; the rows stay on your machine.
Permissions it asks for
- File Read — to access data files and understand their structure
- Shell — to run shell-based data pipelines (awk, jq, sed)
- Sandbox Execute — to run Python or Node scripts in a sandboxed environment
- File Edit — to write output files and processing scripts
Each is permission-gated — Lapu AI asks before it runs.
Just ask
Say it in plain words
No commands to learn. Tell Lapu AI what you want the way you would tell a coworker.
You
Read the CSV at ~/reports/sales-q1.csv, remove all rows where the amount is zero, and save the cleaned result as a new file.
You
Merge all JSON files in the data/ folder into a single CSV, using the 'id' field as the primary key.
You
Parse the nginx access log and generate a summary showing the top 20 URLs by request count with a Python script.
Ready to try this workflow?
Download Lapu AI and run it on your own machine. Free to start — see exactly what it looks like first.
- 1-click uninstall
- Cancel anytime
- Files never leave your computer
FAQ
Common questions
What data formats does Lapu AI support?
Lapu AI can read any text-based format including CSV, TSV, JSON, JSONL, XML, YAML, and plain text logs. For binary formats like Excel (.xlsx), it uses the built-in XLSX skill or command-line tools.
Is my data sent to the cloud for processing?
No. The data processing itself happens locally — via shell commands or sandboxed Python/Node scripts running on your machine. Only structural context (like column names and sample rows) is sent to the AI model to generate the processing logic.
What is sandbox execution?
Lapu AI can run Python and Node scripts in a sandboxed environment on your machine. It automatically installs approved packages if needed, captures output (up to 60KB), and enforces timeouts. This is safer than running raw shell commands for complex data transformations.
Explore more
Related use cases
Document Processing
Process contracts, invoices, and reports without uploading them anywhere. Built-in skills extract, merge, and convert PDF, Word, Excel, and PowerPoint files on your machine.
See how it worksResearchResearch & Extraction
Search thousands of files in seconds. The agent uses Grep, Glob, and File Read to find exactly what you need, extract the relevant pieces, and compile a structured summary.
See how it worksAutomationScheduled Automation
Set it once, let it run. Turn any conversation into a reusable workflow that fires on a schedule — daily reports, recurring cleanups, and routine data pulls happen without you.
See how it works