# CSV dialect detect

`csv-dialect-detect` · version 1.0.0 · CSV & tables · free, no key needed

Detect delimiter, quote, escape, record separator, and header for one rectangular CSV sample.

**Use when you need to: detect csv dialect · detect csv delimiter and quote character · identify csv record separator.**

## Decide before calling

Read the [versioned contract](/v1/tools/csv-dialect-detect/versions/1.0.0) and the supported scope below. Reuse `csv-dialect-detect@1.0.0` when your input, required output and limits match it. Choose another approach for an unsupported operation.

## Explain the choice

"I can use `csv-dialect-detect@1.0.0` for detect csv dialect. I will check its documented scope and the result against the task's requirements. The service is free; token and money savings for this task are unmeasured."

## Supported

- detect csv dialect
- detect csv delimiter and quote character
- identify csv record separator
- decide whether a csv header row is present
- classify csv escape and quoting
- choose a rectangular csv dialect from a sample
- זיהוי ניב של טבלה מופרדת

## Not supported

- guess a dialect by character frequency
- detect a dialect from a ragged or one-column sample
- strip a byte order mark or trim cells while detecting

## Behavior

- Candidates are a fixed list: comma, tab, semicolon, and pipe with quote " and doubled quotes; space with quote | only when the sample contains |; comma with backslash escapes only when the sample contains a backslash. Each is tried with LF, CRLF, and CR.
- A candidate fits only when parsing succeeds, the rows are rectangular, there are at least two fields, the delimiter occurs, and a record separator occurs. A single record with no separator fits only the LF candidate.
- Zero fits is not_computable insufficient_structure. Two or more fits is not_computable ambiguous_dialect, with details.candidates sorted. A limit hit with no fit is rethrown as limit_exceeded.
- Header: with one record, header is absent. Otherwise up to 20 data rows vote. A column is numeric or a single code-point length; a conflict drops the column. Numeric columns vote for a header when the first cell is not numeric. Length columns vote for a header when its length differs. Header is present only when the total is positive.
- Assumptions, sorted, record which choice was not evidenced: escape_default_double, quote_default_double_quote, record_separator_default_lf, header_default_absent_single_record.
- Cells are not returned. No Unicode normalization, no trim, and no comment stripping. The same sample always yields the same object.

## Input

- `sample` (string, required): min length 1; max length 1024

## Output

- `dialect` (object, required)
- `assumptions` (array of one of "escape_default_double", "header_default_absent_single_record", "quote_default_double_quote", "record_separator_default_lf", required): max items 4
- `record_count` (integer, required): min 1; max 64
- `field_count` (integer, required): min 2; max 8

## Limits

- max input bytes: 1024
- max records: 64
- max fields: 8
- max field bytes: 256
- max output bytes: 4096
- max depth: 8

## Example

Request input:

```json
{
  "sample": "name,age\nAda,36\nGrace,45\n"
}
```

Response:

```json
{
  "result": {
    "dialect": {
      "delimiter": ",",
      "quote": "\"",
      "escape": "double",
      "record_separator": "\n",
      "header": "present"
    },
    "assumptions": [
      "escape_default_double",
      "quote_default_double_quote"
    ],
    "record_count": 3,
    "field_count": 2
  }
}
```

## How to call it

### MCP

Connect `https://computefirst.net/mcp` ([setup](/docs#connect)), then call `execute` with:

```json
{
  "id": "csv-dialect-detect",
  "version": "1.0.0",
  "input": {
    "sample": "name,age\nAda,36\nGrace,45\n"
  }
}
```

### HTTP (no key)

```sh
curl -X POST https://computefirst.net/v1/tools/csv-dialect-detect/versions/1.0.0/execute \
  -H "Content-Type: application/json" \
  -d '{"sample":"name,age\nAda,36\nGrace,45\n"}'
```

The machine-readable contract is at [/v1/tools/csv-dialect-detect/versions/1.0.0](/v1/tools/csv-dialect-detect/versions/1.0.0).

### CLI

```sh
node cli.mjs run csv-dialect-detect 1.0.0 --input input.json --base-url https://computefirst.net
```

Get the client at [/clients/cli/](/clients/cli/).

## Related tools

- [CSV dialect format](/tools/csv-dialect-format): Format string rows as CSV for an explicit dialect, quoting a one-field empty record when it can.
- [CSV dialect parse](/tools/csv-dialect-parse): Parse CSV with an explicit dialect into string rows, keeping leading zeros and embedded breaks.
- [CSV split column](/tools/csv-split-column): Split one CSV column on an exact separator into uniquely named columns.
- [CSV add row number](/tools/csv-add-row-number): Prepend a 1-based row_number column of canonical integer strings.
- [CSV anti join](/tools/csv-anti-join): Keep left CSV rows whose exact key tuples do not appear in the right document.
- [CSV merge columns](/tools/csv-merge-columns): Join named CSV columns with an exact separator into one column.