Review · Manatal

Manatal review: it fills in blanks you did not ask it to fill

By Teruaki K. · Published Aug 27, 2026 · Updated Sep 6, 2026

What I actually did with it

I used Manatal for six days, from 22 to 27 August 2026, on its fourteen-day trial. In that time I ran the onboarding, imported a candidate spreadsheet, uploaded two résumés — one detailed, one deliberately generic — created a job with requirements attached, looked at what the AI would and would not tell me, and exported every candidate with all 52 columns switched on.

The candidates are test data I generated. The job was created but left unpublished, so nothing went to a job board and no real applicant was involved. I had spent the previous eleven days doing the same things in a competing product, deliberately, with the identical files — that comparison is a separate article, and this one is about Manatal on its own.

One caveat that shapes everything below. The trial does not appear to run on the $15 Professional plan that a small team would actually buy. There is no label anywhere saying which tier you are on, but the subscription screen shows Enterprise pre-selected, features the pricing page puts above Professional are switched on, and a note reads “By selecting Enterprise you will not have access to Enterprise Plus features such as Groups, SSO and API” — which is the phrasing you use for something a person currently has. So treat every capability below as seen on the trial tier, not as a description of the entry plan.

You cannot sign up with a Gmail address

The first screen rejected my personal address:

We’re currently only accepting sign-ups from work email accounts. If you don’t have a work email, reach out to us at support@manatal.com.

I had already hit the same wall in the other product I was testing, which told me only “Please use a company address.” Both vendors stop you at the door if your company does not own a domain. The difference is that Manatal gives you somewhere to go. Whether writing to that address actually gets you an account, I did not test — what I can report is that one product offers an exit and the other does not.

I signed up with an address on my own domain and it went through without incident.

The onboarding asks which product you are leaving

Manatal asks four questions before it lets you in, and the competing product asked none.

The first is whether this is an Agency Account — recruiting on behalf of external clients — or a Company Account, hiring for yourself. That distinction runs through the whole product, and it is worth noticing on a pricing page too: a per-seat product aimed partly at agencies prices very differently from a flat-rate product aimed at in-house teams. I chose Company Account and Just me.

The second asks where you want to start — career page, sourcing, centralising your data, collaboration, or automation — with a note that “Your choice will define your onboarding experience, but you will still have access to all the features.”

The fourth is the interesting one. It asks what you used before, and if you answer ATS it makes you type the product’s name into a required field. I typed the name of the competitor I had just spent eleven days in. I have no evidence about what happens to that answer, so I will only say what the screen does: it is a required field, and it is the only question in the flow that cannot be answered by picking from a list.

Three concepts that were not in the other product

The sidebar has Departments, Matches and Employees, none of which have an equivalent in the product I came from.

Departments are containers: “Creating departments will allow you to structure your account by creating jobs, contacts and guests under each of those departments.” Employees is where hires land — “All employees will be displayed on this screen once the first hire has been made.”

Matches sounds like an AI feature and is not. “Matches are created when a candidate is added to a job.” It is the join between a person and a role. I had assumed from the name that this was the recommendation engine, and I was wrong; there is nothing AI-flavoured on the screen.

The genuinely different one is Sourcing Hub, an external database of more than six hundred million profiles assembled from public data, which Manatal is careful to keep at arm’s length from your own records: “This feature is entirely separate from your internal database, and no information from your account is included in the external candidate database.” That separation is the right way round. Whether a searchable index of people who never agreed to be indexed belongs in a recruiting tool at all is a larger question than this review, and I am flagging it rather than settling it.

Moving a spreadsheet in

This is where Manatal is at its best, and it is not close.

The importer takes a CSV or JSON up to 10MB and offers 24 fields to map onto: employment status, salary expectations, notice period, nationality, diploma, university, GDPR consent and more. The competitor I had just come from offers four. If you are migrating a spreadsheet that people have actually been maintaining, this is the difference between keeping your data and retyping it.

Three things it gets right that I did not expect.

Your header row does not become a candidate. I imported three people and got three people. The other product turned the header row into a fourth record every single time.

It tells you what it is dropping. Before the import runs it says “Unmapped columns will not be imported to the platform” and shows a count — Importing 5 of 6 columns. You are warned, in numbers, that something is being left behind.

Nothing is auto-assigned. Every column starts blank; a column literally named email is not steered towards the email field. That is more clicks, and it also means nothing gets silently mismapped. I am not going to call either behaviour correct.

Two things it gets wrong.

There is no way to choose the job or stage. Not on the mapping screen, not on the confirmation screen. Candidates arrive in the database and where they sit in your process is something you redo by hand afterwards. The other product let me pick.

The surname is not dropped — it is overwritten. Candidate Name is a single field, and once you map it to first_name it disappears from the choices for last_name. So the surname has nowhere to go. What lands in the record is not an empty surname:

Everyone in my import came out as Delta Delta, Echo Echo, Foxtrot Foxtrot. The surnames in my CSV are nowhere in the account. The warning I praised above is honest as far as it goes — it did tell me five of six columns were being imported — but it does not tell you that the first name will be copied into the surname field.

What the parser added that the document did not contain

I uploaded a résumé to a fresh account and got a candidate record with a first name and a surname in it. Neither name is in the document.

I checked this properly rather than assuming. The PDF is 2,939 bytes, contains no compressed streams and carries no author or producer metadata. Searching the raw bytes of the file for either name returns nothing. Then I did it twice more, minutes apart, on the same account. All three uploads produced the identical pair of names.

I am withholding the name itself, because it reads like a real person’s and I cannot establish that it is not. I am also not going to tell you what caused this, because I do not know. My first instinct — that I was seeing another account’s data — is not well supported by a value that repeats exactly, and I have no second theory I can evidence.

What I can report is the behaviour, and there is one more piece of it that makes the whole thing stranger. When the parser cannot read a name, Manatal stops and shows you a modal:

The candidate’s name is missing for one of the following reasons: 1. The candidate’s name could not be retrieved. 2. The resume parsing is taking longer than expected…

That dialogue is genuinely good design. It is more than the competing product does, which registered a spreadsheet header as a person without comment. But the screen telling me the name could not be retrieved had already written a first name and a surname into the name fields. I cancelled and waited three minutes in case the second explanation applied. Nothing changed.

It is not only names. The three records are identical on thirteen of the fourteen fields in Additional Information. On the fourteenth they differ: one of them has Nationalities set to Japanese, and the other two are empty. There is no nationality in the résumé. The document says Tokyo, Japan as a location, and nothing else. Same file, same route into the product, two minutes apart, and one record acquired a value the other two did not.

Every skill it found scored ten out of ten

The parser also creates skills, and puts a number next to each one.

The number is a proficiency rating on a scale of one to ten — click a tag and you get an editing dialogue with 1 to 10 as buttons, the current value highlighted. It is the field a human fills in after talking to someone.

Across the two candidates I inspected, the parser had created fourteen skills, and set every single one of them to 10. Nobody had reviewed these candidates. Nobody had spoken to them. The skills themselves are accurate — the six it pulled off one résumé match the document exactly, and it did not invent a skill the way the other product I tested did — but the ratings attached to them are a top mark that no one awarded.

I have not seen a value other than 10, and I am not going to claim the default is always 10 on the strength of two candidates. What I will say is that in everything I observed, the score was maximum, and nothing on the screen distinguishes a rating the parser generated from one a recruiter entered.

The AI answers when asked, and says when it does not know

The AI here works in a way I ended up liking, and it is worth being specific about why, because the two products I compared take opposite approaches and neither is obviously right.

Manatal’s Copilot is on by default, but it runs only when you press Ask AI and pick a question. Every answer carries the line “Beta version, results may be incomplete.” A Sources button lists the data types the assistant can see — candidate details and emails, job details, match details, department details, notes. That is a list of what it can read, not a citation for the answer in front of you, and it should not be mistaken for one.

Then the part that actually mattered. I asked it to summarise a candidate who had no notes on file. It ended its answer with:

No notes have been added to this profile.

That sounds like nothing. It is the whole thing. The competing product, asked the same question about a comparably empty record, produced a dated timeline of events that had not happened. Manatal reported the blank as a blank.

The summary itself was accurate too — account numbers, revenue figures, team size, all matching the résumé. And the experience figure it quoted was right, which brings me to the one that is harder to write up.

The number that is right and is not saved

Asked about experience, Manatal answered “approximately 12 years”, which is correct for the document I gave it.

The structured field tells a different story. Years of Experience reads N/A on the candidate screen, and in the 52-column export I took on 27 August it is empty for all seven candidates in the account — while the AI Recommendations panel, at the same time, on the same account, displayed 12 years of experience under a candidate’s name.

So the figure exists on screen and nowhere else. It is not in the record and it will not come out in a file. Whatever produced it, you cannot keep it, quote it back, or show it to anyone later.

And it is not always right. On a second, deliberately generic résumé describing nine years, Manatal answered “approximately 5 years” — four years low, a larger error than anything the competing product produced on the same two documents. I would not describe this parser as accurate. I would describe it as exact once, badly wrong once, and unwilling to write down either answer.

Where the credits actually are

Manatal’s pricing page says the AI is included in every plan, and as far as I could tell that is true: the AI settings screen shows no credit balance, no cap and no usage meter. After eleven days in a product that bills AI screening in credits you buy separately, this is a real relief.

The metering has not vanished, though — it has moved. Sourcing Hub runs on search credits: “Each search credit gives you access to up to 100 profiles”, with a balance reading 100 of 100 search credits left, resets in 9 days. Alongside it, the same panel reads Profile Balance: Up to 1,000 profiles accessible with your remaining search credits, which does not divide into the first sentence the way I would expect. I am recording that as something I saw rather than diagnosing it. Automations are capped too, at 0 / 30.

I have since measured what a search credit actually buys, and what a sourced record arrives carrying. If you are an agency rather than an employer, that and the client-separation side are covered in what an ATS costs a small recruitment agency.

There is also an AI Interviewer, on its own screen, behind its own toggle, with no price anywhere on the page — you find out what it costs by enabling it. The description carries a line worth reading twice: “The availability of recordings older than one year may change in the future.”

Creating a job, and why you cannot post it

Job creation is about a dozen fields with three required: position name, department, pipeline. Skills and education are not ordinary fields here; they live in a separate Job Requirements block, marked Required or Preferred, and that block is what the recommendation engine reads.

The best thing about this screen is what happens when you submit. The job saves as NOT PUBLISHED. You do not have to find a “save as draft” button, and there is no way to push a role live by accident.

The worst thing is what happens next. Job Boards lists 36 free destinations — Indeed, LinkedIn, Monster, Google for Jobs and the rest — and every one of them is locked:

The feature is not available during the free trial. Please subscribe to Manatal and complete your account verification to use it.

No board is on by default, which is the safe setting. But no price is shown anywhere, for any board, and premium posting is gated behind enabling a career page I chose not to turn on. So the thing you would most want to evaluate during a trial — where the job goes, and what that costs — is the one thing the trial will not show you.

What comes back out

Turn on all 52 columns and you can export effectively everything: names, contact details, salary expectations, GDPR consent, the candidate description. That is a genuinely generous export, and every column was available to me with no lock icons and no upgrade prompts. What those columns actually contain is worth counting separately: thirty of the fifty-two came out empty on every row, and the full tally is here.

Four things to know before you rely on it.

The export columns are the list’s display columns, and the default selection of ten hides email address, first name and surname. It also includes Notice Period, Current Salary and Expected Salary, which were empty for all seven of my candidates. If you export without opening the column editor you get a file that looks complete and is missing what you came for. I misread this myself before I worked out what was happening.

No AI output is in there. I searched the header row for score, match, recommendation and AI: there is no score, no breakdown, no reasoning in any of the 52 columns. Whatever the recommendation engine concluded stays inside Manatal.

Candidate Location cannot be imported. It is one of the 52 columns you can export and it is not one of the 24 you can map on the way in, so an address you take out is an address you cannot put back.

Pressing Export does not produce a file. The export runs in the background and posts a notice reading “Your export of candidates has been successfully completed. Please use the following link to download it here.” The file only exists once you click that here link. I lost a session to this: the banner said the export had succeeded, and my downloads folder was empty.

There is no free plan underneath this

Manatal has three published tiers — Professional at $15 per seat per month billed annually or $19 monthly, Enterprise at $35 or $39, Enterprise Plus at $55 or $59 — and nothing below them. The annual toggle is badged 10% off; the actual discounts are 21%, 10% and 7% depending on tier.

The trial takes no card, which I appreciated, and which means nobody gets charged by accident. It also means the account has no payment relationship to fall back on. When the fourteen days end there is no reduced tier for your data to land in, and the interface never tells you the trial’s end date — only Free trial 14 days left, counting down.

There is no explanation anywhere of what happens when it reaches zero. I looked. On 5 September it reached zero, and the whole product became a checkout — no export, no delete, three links.

Who should use Manatal

Take it if you are ten people or fewer and you have real data to move. The importer is the best I have used, the per-seat price is genuinely low at small headcounts, and the job creation flow will not distribute anything without being told to. If you are an agency billing per recruiter, the whole product is shaped for you in a way the flat-rate alternatives are not.

Do not take it if you need to stop and start. No free plan means no gap between hires, no pausing for a quiet quarter, no coming back to an account you left. That is a bigger constraint than it sounds like for a small company, and it is the single thing I would weigh hardest.

Do not rely on the parser being exact. It reads structure well — location, phone, diploma, university, skills — and then writes values into fields the document never filled: a name, a nationality, fourteen top marks. Everything it produces is a draft that needs a person to look at it, which is true of every tool in this category and more visibly true here.

And check the record, not the screen. The figure a screen shows you may not be in the record underneath it, and if you ever have to explain a hiring decision, the export is what you will have.

Disclosure: the link to Manatal in this review is an affiliate link, which means I may earn a commission if you sign up through it. It costs you nothing extra, it does not change the rating, and I do not accept payment in exchange for a favourable review.