Automate
From document
to a reviewed record
For recruiters who receive CVs and client job descriptions as files, e-mail text or links. Upload or paste the document, or give the link to a job posting, and Jenny reads it and prefills a candidate or job for you to check. Nothing is saved until you do, and facts the document does not state stay blank.
The problem
Retyping a CV is where the errors start
A CV arrives as a PDF, a client brief as the body of an e-mail or a link to a posting. Each one has to become a record before you can search, match or present anything, and copying the same facts from a document into a system by hand is slow and easy to get wrong.
Handing the job to a tool that guesses is no better. A salary nobody mentioned, “fluent” where the CV said “basic”, or an example read as a mandatory requirement misleads every match that follows. Extraction only helps if you can see what came from the document and what is still unknown.
What changes
- One upload, paste or link turns a CV or job description into a prefilled form instead of an evening of retyping.
- Facts the document does not state stay blank or “Not stated”, so a gap is visible rather than filled with a guess.
- Values read by AI are labelled “Inferred — confirm” on the profile until someone has confirmed them.
- In a job description, alternatives such as “SAP or Oracle”, examples and real must-haves are kept apart, so matching starts from the right requirements.
- A saved record is ready for the next step: finding jobs for the candidate, or searching candidates for the job.
CV review · Candidate A
Illustrative — fictional data- Inferred — confirm
Current title
Financial Controller
Read from the CV
- Inferred — confirm
Languages
Dutch, English
Read from the CV
- Recorded
Experience
2017–present · logistics
Corrected by you
- Missing
Salary expectation
Not stated in the CV
Left blank
- Missing
Notice period
Not stated in the CV
Ask the candidate
How it works
01
Add the document
Upload a CV or job description as PDF, DOCX or plain text (up to 4 MB and 40 pages), paste the text, or give the link to a public job posting.
02
Jenny reads it
The text is read on the server and AI extraction proposes the fields: title, skills, experience, languages and location for a candidate; requirements, location and conditions for a job.
03
Check the form
The prefilled form shows what was found and leaves the rest blank. Correct, add or remove anything before you save.
04
Save once
Nothing is stored until you save, and the record is created a single time. AI-read values that nobody has confirmed stay labelled on the profile.
05
Take the next step
From a candidate, look for matching jobs; from a job, start a candidate search with the requirements you checked.
Worked example: A CV and a client brief on the same morning
Illustrative — fictional data01 · Upload the CV
Candidate A’s CV arrives as a two-page PDF. You upload it from the Candidates page, and the text is read on the server.
02 · Review the profile
The form is prefilled: Financial Controller, eight years in logistics, Dutch and English, based near Mechelen. Salary and notice period are not stated, so they stay blank.
03 · Save and match
You correct one job title and save. The profile lists salary and notice period as missing, and from there you look for matching jobs.
04 · Paste the brief
A client contact e-mails a brief for a Senior Financial Controller at a logistics company in Antwerp. You paste the text; Jenny keeps the must-haves, “SAP or Oracle” and the examples apart.
05 · Check and save
The budget is not mentioned, so it stays unknown. You confirm the requirements and save the job, ready for a candidate search.
06 · Later, by mailbox
In development, not available: the same CV landing in a connected mailbox folder would wait in a review queue, or be linked to Candidate A by itself only on a decisive match.
You provide
- A CV as PDF, DOCX or plain text with selectable text, up to 4 MB and 40 pages, or the pasted text
- A job description as a file, pasted text or a link to a public job posting
- Your corrections and confirmation on the prefilled form
Jenny produces
- A candidate or job form prefilled from the document, with unstated facts left blank
- A saved record once you confirm the form, created exactly once
- A label on each profile fact: recorded and verified, recorded, “Inferred — confirm” or missing
- A flag on the profile when the CV is undated or more than three years old, asking you to confirm the current role and preferences
Sources and their status
| File upload | Available | PDF, DOCX or plain text up to 4 MB and 40 pages, read on the server. Scanned documents without selectable text cannot be read. |
|---|---|---|
| Pasted text | Available | A CV or job description pasted straight from an e-mail or another document. |
| Link to a job posting | Available | A public job page, LinkedIn job links included, read into a job draft. LinkedIn profile and company pages are refused. |
| CV library | Available | Up to 200 authorised CVs (PDF or DOCX) per workspace, searched when you look for candidates for a job. |
| Connected mailbox folders | In development | CVs and job descriptions from Gmail or Microsoft 365 folders you select, into a review queue. Not available. |
| Drag and drop | Available | Drop a CV or job description anywhere in your workspace, or into the chat where it is available. Jenny first checks on her own server — without AI and without saving anything — whether it is a CV or a job description, then opens the right setup. A job description’s details are extracted at once (if it also looks like a CV, you choose first); a CV’s only after you confirm you may use it. A record is created only when you click Save or Find matches. |
What it depends on
- Documents with selectable text: PDF, DOCX or plain text up to 4 MB and 40 pages (no OCR)
- AI extraction by Claude, within your workspace’s daily AI limits; where AI extraction is not switched on, a CV is read with simpler local rules, labelled as such
- For job links: a public page that can be opened without signing in
- For mailbox intake, once released: a connected Gmail or Microsoft 365 mailbox and someone allowed to switch intake on
Approval and privacy controls
- Nothing from a document is saved until you review and save the form.
- Unknowns stay unknown: an unstated salary, notice period or language level is left blank, never estimated.
- AI-extracted values that nobody has confirmed are labelled “Inferred — confirm” on the profile.
- The same CV file added to the CV library twice is recognised and stored once.
- Only public job pages are read from a link; links to LinkedIn profiles and company pages are refused.
Current limitations
- There is no OCR: scanned PDFs and photos of a CV cannot be read, so paste the text instead.
- Files are limited to 4 MB and 40 pages, and the CV library holds up to 200 documents per workspace.
- Every document is added by a person today. Reading CVs and job descriptions from a mailbox is in development and not available.
- Applications through the careers page do not include a CV upload.
- An unusual layout can be misread, which is why every value is shown for review before anything is saved.
Questions about cv and job-description intake
Which files can Jenny read?
PDF, DOCX and plain-text files up to 4 MB and 40 pages, as long as the text can be selected. Scanned documents and photos are not read because there is no OCR; copy the text from the original and paste it instead.
What happens to information the CV does not mention?
It stays blank. A salary, availability or language level the document does not state is left as “Not stated”, and the profile lists it as missing so you know what to ask. Jenny does not contact the candidate to fill the gap.
Can I add a job description from a link?
Yes, for public job pages, LinkedIn job postings included. Jenny reads the page into a job draft that you review in the same way as an uploaded file. Links to LinkedIn profiles or company pages are refused.
How can I tell which values came from AI?
On the candidate profile, values read by AI extraction that nobody has confirmed are labelled “Inferred — confirm”. Facts you recorded or verified are labelled as such, and missing facts are listed as missing.
Can Jenny pick up CVs from my inbox?
Not yet. Intake from a connected Gmail or Microsoft 365 mailbox is in development: it would read only the folders you select, put each CV or job description in a review queue and link a CV by itself only on a decisive match. Until then, upload or paste each document; for an existing database, the CSV import is the better route.
Related features
- Candidate-to-job matching
Start from a candidate and find current postings that fit, with reasons for each.
Early access - Job-to-candidate matching
From a job description to an explained shortlist — evidence, gaps and unknowns included.
Available - E-mail integration
In development: mailbox folders you select, a review queue and reply drafts on request.
In development - CRM/ATS integrations
Work alongside your CRM or ATS: CSV import today, a native connector on request.
Early access
See it on your own vacancies and candidates
A short call with the founder: we look at how you work today and show the workflows that fit.