Spreadsheet Column Mapper

Q: Does the AI read my spreadsheet rows?

No. Optional AI mapping receives only source and target header names. The table rows remain in the deterministic browser workflow.

Q: Can two source columns map to one target?

No. This workflow enforces one-to-one mappings and rejects duplicate target use instead of silently overwriting data.

Q: What happens to unmapped columns?

Unmapped target fields are exported blank. Unmapped source fields are excluded from the mapped table and listed in the manifest.

Q: Can I restore the mapping after refresh?

Normally yes when persistent browser storage is available. Clearing site data, private browsing, or storage eviction can remove saved files and model weights.

Map CSV or Excel columns to a target schema in your browser. Exact normalized matches are suggested deterministically; optional local AI sees headers only, and every one-to-one mapping must be confirmed before export.

Saved sessions

Drop a source CSV, TSV, or XLSX file here

Rows stay in this browser. Legacy .xls files are not supported.

Local AI runtime

Checking this device

BackendWASM

Device profile—

Adaptive limits—

Writing model

Qwen 0.5B · standard

Transcription model

Whisper Tiny · fast

Files, prompts, and results stay in this browser. Browser storage can still be cleared by you, private mode, or browser cleanup.

Private, browser-local schema mapping

The source table, headers, mappings, mapped output, and manifest stay in your browser and are excluded from BaseToolbox analytics. Optional local AI receives only the displayed header lists.

Current limitations

CSV, TSV, and modern XLSX are supported; legacy .xls, macros, and desktop-Excel formula recalculation are not.
Deterministic suggestions require an unambiguous normalized header match. Similar meanings need manual review or optional local-AI suggestions.
The tool does not transform units, data types, currencies, or values; it only reorders and renames fields according to confirmed mappings.
Browser storage may be cleared or reclaimed, and private browsing may not retain source files, mappings, or downloaded model weights.

How column mapping works

Import data and define targets

Load the source table and enter the exact target column list or paste a target schema example.

Review deterministic and optional AI suggestions

Exact normalized matches appear first. The shared local model can suggest remaining header pairs without seeing row values.

Confirm and export

One-to-one validation creates a mapped table, an unmapped-field report, and a JSON manifest alongside CSV or XLSX export.