A folder of PDFs in. A working pipeline out.
AI Screen reads every resume that reaches you — a PDF, a Word file, a phone photo of a printed resume — pulls out a structured candidate record, checks it against three gates, matches it to the right open role, and routes it. Around 2.4 seconds each, and there is no limit on how many.
- No credit card
- No per-resume cost
- Nothing is ever discarded silently
Part of the Tanca ecosystem
- Tanca
- Skillify
- Glinks
- GCafe
- GHR
- GWork
The most skilled person on your team is doing data entry.
Someone on your team opens a PDF, reads it, types a name into a form, types a phone number, checks whether this person already applied last quarter, decides which opening they might fit, and moves on to the next one. Three hundred times per opening.
None of that is recruiting. It’s transcription, and it’s the reason your first shortlist takes two weeks.
Scan as many resumes as you like. There is no counter.
This is the feature most systems in this market meter. We don’t — and here is the exact boundary between this and the thing that does have an allowance.
Unlimited, permanently
- Resumes scanned
- Pages per resume
- Documents read from photos
- Field extraction
- Duplicate checking
- Restricted-list checking
- Consent capture
- Position matching
- Match scoring at intake
- Routing
- Batch uploads
- Trace logs
- Re-scans
- Source connections
Draws on your AI allowance
The match score AI Screen produces at intake is not AI Score and does not touch your allowance. Intake scoring answers “does this person fit this opening at all?” AI Score answers “at what level, on which skill, against which framework?”
See how the allowance works →Nothing about this engine is metered. If you drop four thousand resumes in on a Monday, four thousand get read.
Including the ones photographed on a phone.
Frontline hiring does not arrive as clean PDFs. It arrives as a photo of a printed resume taken at a store counter, a screenshot forwarded by text, a handwritten application form, a Word file made from a template downloaded in 2019.
AI Screen reads all of it, in English and in Spanish, and it keeps accented characters and non-English names intact instead of flattening them into three different spellings the way parsers built only for ASCII do.
When it can’t read something confidently, it says so and sends the record to your review queue with the uncertain fields marked. It never guesses silently.
- PDF, DOCX, JPG, PNG, HEIC
- English and Spanish, accents intact
- Low-confidence fields flagged, not guessed
exported from any resume builder
DOCX
including old templates
Phone photo
printed resume, taken at a counter
Screenshot
forwarded by text message
Scanned form
handwritten application
Email body
no attachment at all
Three checks, named, with their results shown.
Not a black box that returns a verdict. Three specific questions, each with an answer you can see on the record.
Duplicate check
Has this person applied to you before, under a different email or a different spelling of their name? If so, the records merge instead of splitting your history in two.
Restricted list
Is this someone your company has decided not to process? Checked at intake, before anyone spends time on the file.
Consent and disclosure
Did this person agree to their data being handled, and when exactly? The record a CCPA or GDPR request asks you to produce. Captured at intake with a timestamp, so you’re never reconstructing it later.
And when a gate doesn’t pass, you see that too.
A resume typed in by hand has no consent record, and the product says so rather than pretending otherwise. That flag follows the record everywhere it goes, including into your talent pool.
Move it and the split updates before you commit.
The number that decides what’s automatic is yours.
Set the auto-approve threshold — the default is 85. A resume that matches exactly one open role above that number goes straight into the pipeline. Everything else goes to your review queue.
Raise it and more comes to you. Lower it and more flows automatically. Change it any time, per position if you want. The engine doesn’t have an opinion about where the line should be.
Ambiguity always goes to a person. If a resume matches two openings equally well, no threshold makes that automatic — it lands in review with both suggestions attached.
- Default 85, adjustable any time
- Per position if you need it
- Ties always go to a human
Every resume ends up somewhere you can find it.
There is no bin. Four outcomes, each visible on the record and each reversible by a person. The engine sorts; it never decides who is worth looking at.
To pipeline
Matched one open role above your threshold. Already in the right stage, ready to work.
134 todayTo review
Below the threshold, or matched more than one role, or a field the engine couldn’t read confidently. A person decides.
5 todayTo talent pool
No open role fits today. Kept, searchable, and surfaced by the Re-engagement Agent when something opens.
—Blocked
Matched your restricted list, or is a confirmed duplicate of an existing record. Visible, with the reason.
3 todayWhat the engine will never do.
It will never delete a resume, never hide a candidate from your view, and never mark someone as unsuitable. “Below your threshold” means a person should look at this, not this person isn’t good enough — and the record says exactly that, in those words, on screen.
Drop in a folder. Review what needs you, approve the rest together.
The queue is grouped by batch, so a Monday-morning upload stays one job instead of forty notifications.
Blocked records stay visible, with the reason attached.
What happened to this resume, in order, with times.
Open any record and there’s a log: what the engine read, which gates ran and what they returned, which positions it considered, what it scored, where it routed, and when each step happened.
This matters twice. Practically — when a candidate says they applied and nobody can find them, the answer is on one screen. And formally — when someone asks how a decision was made, you have the sequence rather than a recollection.
The log is read-only and it isn’t editable by anyone, including us.
- Per-resume, timestamped
- Read-only, not editable
- Exportable for audit
The log shows where consent came from, not just that it exists.
The engine is the same. The workflow isn’t.
When the position is already known, nothing needs your attention. When it isn’t, you get a queue.
From your careers page
The applicant picked the role themselves, so there’s nothing to guess. The engine reads, gates, scores, and files them in the right pipeline stage. No notification, because nothing needs you.
This screen is a history log. You’d only open it to check something.
From an upload or a folder
Nobody told the engine which role these people are for, so it proposes. Above your threshold with one clear match, it files them. Otherwise it asks.
This screen is a queue. You’d open it once per batch.
The front door to everything else.
Everything downstream is only as good as what comes through the door — which is why this is the part we refuse to meter.
Measured inside the product.
Food retail chain · 22 stores
Hours of data entry removed per 300-resume opening.
Logistics company · 3 southern warehouses
Time from a resume arriving to it appearing in a pipeline stage.
Admissions center · telesales
Duplicate records eliminated in the first batch.
You’ll see the same numbers for your own account from the first batch.
Questions people ask about scanning.
Is resume scanning really unlimited?
Yes. No per-resume charge, no credit pack, no monthly cap. AI Score — the deeper scoring against a competency framework — is the feature that carries an allowance.
Can it read a photo of a printed resume?
Yes, including photos taken on a phone at a counter and screenshots forwarded through a messaging app. Fields it can’t read confidently are flagged rather than guessed.
How good is it with Spanish and non-English names?
It reads Spanish resumes natively and keeps accented characters intact, which parsers built only for ASCII silently drop. Names come through as written.
What happens to resumes that don’t match any open role?
They go to your talent pool with their gates and consent status attached, and the Re-engagement Agent surfaces them when a matching role opens.
Does it ever reject someone automatically?
No. Below your threshold means it goes to a person, not that it goes away. The only records that are blocked are restricted-list matches and confirmed duplicates, and both stay visible with the reason.
Can we change the automatic threshold?
Yes, any time, and per position if you want. The default is 85.
What if a resume has no consent record?
The record shows that plainly, with the reason, and the flag follows it everywhere — including into your talent pool. You always know what you may and may not do with it.
Can we see what the engine did?
Every resume has a read-only trace log with timestamps for each step, and it’s exportable.
Drop in the folder you’ve been avoiding.
Every resume in it gets read, checked, scored and filed — and none of it costs anything.
Start free- No credit card
- No per-resume cost
- Nothing discarded silently