CV dates
Every date comes back as {raw, value, precision}. The value is never
more precise than the CV: a year stays a year, a season stays a season. Code does the parsing,
not a model, so "2019" can't turn into "January 2019".
| In the CV | value | precision |
|---|---|---|
| 2018 | "2018" | year |
| 03/2022 | "2022-03" | month |
| fev. 2022 | "2022-02" | month |
| Spring 2020 | "2020-spring" | season |
| 14.03.2020 | "2020-03-14" | day |
| Present / Presente (as an end) | null, is_current: true | null |
Ambiguity
- Day/month order (03/04/2021) is resolved from the CV's language and location when possible; otherwise only the year is kept and the date gets a
DATE_AMBIGUOUSwarning. - Year-only dates get
DATE_YEAR_ONLY, because any duration calculated from them is approximate.
Try it
Get a free API key (50 parses a month, no card), then:
curl https://api.verbatimcv.com/v1/parse \
-H "Authorization: Bearer $VERBATIMCV_API_KEY" \
-F file=@cv.pdf
Agents can use the parse_resume tool instead: uvx verbatimcv-mcp. See the
quickstart for Python, JavaScript and MCP setup.