Excel Silently Auto-Formats (and Corrupts) Your Data.
We Convert 1M Rows in 10s Without Corruption.
Excel auto-formats 000123 → 123, phone numbers → scientific notation, dates → serial numbers. Force Text Mode preserves your data exactly as-is. Convert Excel to CSV/JSON/TSV at 100,000 rows/second — entirely in your browser.
Excel "Save As CSV" Destroys Your Data
It happens silently. No warning, no error. You open the CSV in Excel, everything looks fine. Then your downstream system rejects the import — because 000123 became 123 three steps ago.
| Input Data | Excel "Save As CSV" | SplitForge Force Text Mode |
|---|---|---|
| 000123 (SKU) | 123 | 000123 |
| 5551234567 (phone) | 5.55E+09 | 5551234567 |
| 01234 (zip code) | 1234 | 01234 |
| 3-7-2024 (date text) | 43897 | 3-7-2024 |
| 1.23E+15 (sci notation) | 1230000000000000 | 1.23E+15 |
What Does This Tool Do?
Converts Excel files (.xlsx, .xls) to CSV, JSON, TSV — or any of those formats back to Excel — without corrupting your data. Unlike Excel's built-in "Save As", it uses Force Text Mode to prevent auto-formatting that destroys IDs, phone numbers, zip codes, and SKUs. Handles 1M+ rows, multi-sheet workbooks, custom delimiters, and files up to 2GB. All processing happens in your browser — your file contents are never uploaded.
Features That Actually Solve Problems
Force Text Mode
On by defaultStops Excel auto-formatting. Keeps 000123 as 000123, phone numbers as text (not 5.55E+09), dates unchanged. The #1 reason people switch from Excel Save As.
1M+ Row Handling
>Excel limitExceeds Excel's 1,048,576 row limit for exports. Converts 1M rows in ~10 seconds at 100K rows/second. Streaming architecture prevents memory crashes.
Multi-Sheet ZIP Export
50 sheets → 1 clickConvert all sheets at once. Output: ZIP with individual CSV/JSON/TSV per sheet. No more one-by-one manual conversion.
Custom Delimiters
Any characterComma, semicolon, pipe, tab, or any custom character sequence. Auto-detects input delimiter with 90%+ confidence. Covers European CSV (semicolons) out of the box.
Skip Header Rows
Skip rows 1–100Ignore metadata rows at the top. Start conversion from actual data at row N. Useful for Excel reports with title rows, date generated, filter criteria above the data.
Remove Empty Rows/Columns
50–90% size reductionAuto-remove trailing blank rows and columns. Excel exports often include hundreds of trailing blanks that bloat file size. Reduces output size 50–90% for typical reports.
JSON & TSV Output
4 formats totalExport to JSON (with optional pretty-print) or TSV. First row becomes JSON keys; each data row becomes an object. Perfect for APIs, databases, BigQuery, and Redshift.
100% Private — Zero Uploads
Client-side onlyAll processing in-browser using Web Workers. Your files never leave your device. No data is sent to any server. Designed to avoid server transmission of PHI and PII.
Files Up to 2GB
Browser-memory boundUp to 2GB max — practical limit depends on available RAM. 16GB+ recommended for 1M+ row files. Streaming architecture processes data in chunks to stay within browser memory limits.
How It Compares
| Feature | Excel Save As | Online Converters | SplitForge |
|---|---|---|---|
| Force Text Mode (no data corruption) | Always corrupts | Inconsistent | On by default |
| 1M+ row support | 1,048,576 limit | Usually 10–100MB | No export limit |
| Multi-sheet ZIP export | Active sheet only | Single sheet | All sheets → ZIP |
| Privacy (no data upload) | Local only | Uploads to their server | Client-side only |
| JSON / TSV output | No JSON support | Limited | CSV, JSON, TSV, Excel |
| Custom delimiters | Comma or tab only | Usually comma only | Any character |
| Skip header rows | No | Rarely | Skip rows 1–100 |
| 2GB file support | Crashes often | 10–100MB limit | Up to 2GB |
Real-World Examples
E-Commerce: Product Catalog Export
250,000-SKU product catalog in Excel (3 sheets: Products, Inventory, Pricing)
Finance: Transaction History Export
Bank exporting 1.2M transaction records for fraud analysis
Healthcare: Patient Record Migration
80,000 patient records from Excel to new EHR system (requires JSON)
Which Tool Fits Your Situation?
Edge Cases We Handle
The real-world messiness that breaks basic converters.
Large Numbers (15+ Digits) — Precision Preservation
Leading Zeros — Force Text Mode Required
Multi-Sheet Workbooks with Duplicate Sheet Names
Files Exceeding Excel's 1,048,576 Row Limit
Encoding Issues — UTF-8, UTF-16, ISO-8859-1
Excel Error Cells (#REF!, #VALUE!, #N/A)
ROI Calculator
Baseline: ~15 minutes per file fixing Excel corruption errors (leading zeros re-added, scientific notation corrected, dates fixed). SplitForge: ~10 seconds per file with Force Text Mode — zero post-conversion fixes.
1 Million Rows in ~10 Seconds
Browser-based conversion at scale — no uploads, no queues, no waiting for server round-trips.
Used in Production — Across Industries
Session analytics and tool usage logs.
Why We Built This
SplitForge started after repeatedly watching Excel silently corrupt SKU and financial data in production systems. Warehouse imports failing on "000123" turning to "123." Phone numbers converted to scientific notation. Dates becoming serial numbers nobody asked for. The tools people were using to "fix" the problem required uploading sensitive business data to random servers — a compliance risk nobody talked about openly. We built a converter that runs entirely in the browser, prevents corruption by default, and never asks for an upload. The privacy moat and the performance moat are the same thing.
Honest Limitations: Where SplitForge Falls Short
No tool is perfect for every use case. Here's where Python pandas or a server-side ETL tool might be a better choice, and the real limitations of our browser-based architecture.
Maximum practical file size is ~2GB. Processing 1M rows allocates ~400–600MB of browser memory. Files above this risk out-of-memory crashes depending on column count and RAM available.
SplitForge is a browser tool — no REST API, CLI, or pipeline integration. Cannot run on a schedule, in CI/CD, or be triggered programmatically.
Converts one file at a time. No bulk-batch processing across multiple files in a single operation (multi-sheet ZIP export is supported within one file).
This tool converts formats — it does not merge multiple files. If your workflow requires combining output files, you need a separate step.
When Python pandas or Server-Side ETL Is the Better Choice
Who Should NOT Use SplitForge
Frequently Asked Questions
What file formats are supported?
What is Force Text Mode and why do I need it?
What are the file size limits?
Can it handle files beyond Excel's row limit?
How do I convert multiple sheets at once?
What does "Skip Header Rows" do?
When should I remove empty rows/columns?
What is Pretty-Print JSON?
Is my data safe? Does it get uploaded?
What browsers are supported?
Does SplitForge have any limitations?
SplitForge vs Python pandas — when should I use which?
Is SplitForge free? Are there usage limits?
Related guides
Stop Letting Excel Corrupt Your Data
Force Text Mode. 1M rows in 10 seconds. Multi-sheet ZIP export. File contents never leave your browser.
Start converting in 30 seconds — no account, no install, no upload.
Also try: Excel Splitter · Excel Cleaner · Sheet Merger · Data Cleaner
Maintained by the SplitForge engineering team · Last updated February 2026