DeerX HRMS

HomePricingQuestions
Talk to us
DeerX HRMS · HR, payroll and attendance in one place

Run payroll
without chasing anyone.

Set the rules once. Every month applies them the same way — and shows its working.

Start free — 15 daysBook a walkthroughWatch payroll compute
A schema per company
39 modules, one workspace
01 / 06

People

One record per person, and everything else reads from it.

One row per person
Every change dated
Nothing retyped
WHAT IT DOES

One record. Payroll, attendance and leave all read the same row, so a transfer is one change, not four.

Employee records
Org chart
Lifecycle events
Documents
E-signature
Checklists
+1 more

What it won't do · No drag-and-drop org chart — it comes from the manager field, so it cannot disagree with payroll.

02 / 06

Time

Attendance, leave and shifts that agree with each other.

Punch → approve → pay
Grace from the shift
Every edit audited
WHAT IT DOES

Punches build attendance, every correction is audited, and an approved day is a fact payroll reads.

Attendance
Device ingest
Shifts & rosters
Leave
Holidays & events
Timesheets
+1 more

What it won't do · No geofencing or selfie punches — proving somebody stood in a spot is a different product.

03 / 06

Money

Payroll you can show your working for.

Effective-dated rates
Every line traceable
Shrawan → Ashad
WHAT IT DOES

Every payslip line names where it came from, and a run that cannot compute somebody stops and says who.

Salary structures
Formula engine
Tax slabs
Payroll runs
Payslips
Loans & advances
+3 more

What it won't do · No money moves from here. Payroll writes the bank instruction; your bank makes the payment.

04 / 06

Growth

Hiring, goals and reviews on the same records.

Offer becomes employee
Reviews read the org
Training on the record
WHAT IT DOES

A hire becomes an employee in one step, and reviews read the reporting lines already in People.

Goals & OKRs
Review cycles
Training
Surveys
Recruitment
Careers page

What it won't do · No CV parsing or scoring — ranking people by keyword match is a bias engine with a progress bar.

05 / 06

Comms

The messages that belong to the work, beside the work.

One directory
Tickets with owners
Mail kept per workspace
WHAT IT DOES

Announcements, chat and tickets run on the same identities — no second directory to keep in step.

Chat
Mail
Helpdesk
Announcements
Notifications

What it won't do · Not a chat product. Enough to ask about a record without leaving it, and no more.

06 / 06

Business

Clients, deals and invoices, for the companies that also sell.

Real staff on projects
Client desk kept apart
Invoices sum from lines
WHAT IT DOES

Clients, deals, projects and invoices share the workspace with the people doing the work.

CRM
Invoicing
Reports
Import & export
Operator console

What it won't do · No payment gateway. An invoice records what is owed; collecting it is your bank’s job.

Built in the terms payroll here is actually run in

BIKRAM SAMBATSHRAWAN → ASHADSSFPROVIDENT FUNDCITNEPALI TAX SLABSLAKH GROUPINGTDS
WHAT IT DOES

Three things, and they share one record.

Most companies buy these separately and spend the year reconciling them. Here they are one workspace, so a transfer in March is one change rather than four.

People

One record per person, and everything else reads from it.

The same employee exists three times — in a spreadsheet, in the payroll file, and in whoever knows they transferred in March. When the three disagree, payroll is wrong and nobody can say when it started.

7 capabilities

Discover

Time

Attendance, leave and shifts that agree with each other.

Leave is approved in email, attendance comes off a device, and the two only meet when payroll runs and somebody is docked for a day they were granted.

7 capabilities

Discover

Money

Payroll you can show your working for.

Payroll is a spreadsheet one person understands. When somebody asks why their tax changed, the answer takes an afternoon and rests on trusting that no formula was dragged one row too far.

9 capabilities

Discover

02
Where the month goes

Your month, side by side.

Pick a day. Left: HR in spreadsheets and a chat thread. Right: the same day with the answers already in one place.

a month in the lifeday 01 / 31

Drag through the month

Someone starts today

HOW IT GOES NOW

A photocopied ID goes in a drawer. The offer letter lives in a chat thread. IT learns at 10:40 that a laptop was needed at 9:00.

HOW IT GOES HERE

Marking the candidate hired opens an onboarding checklist. IT, finance and the manager each get their own task with their own due date.

backed by  checklists · recruitment → employees
← drag the handle through the month →
03
See the product

This is the screen, not a drawing of it.

Real captures from a running workspace. The Nepali month names, the lakh grouping and the statutory columns are not a localisation layer — they are what the engine computes in.

yourcompany.deerx.app
Payroll in DeerX HRMS

Periods are Shrawan and Ashad, not July and August. Figures group as 8,21,417 — lakhs, not thousands. Nothing here was translated after the fact.

04
The six domains

Six domains, one set of records underneath.

Each one says what it will not do, as well as what it will.

People

One record per person, and everything else reads from it.

Employee records · Org chart · Lifecycle events · Documents · E-signature

Learn more

Time

Attendance, leave and shifts that agree with each other.

Attendance · Device ingest · Shifts & rosters · Leave · Holidays & events

Learn more

Money

Payroll you can show your working for.

Salary structures · Formula engine · Tax slabs · Payroll runs · Payslips

Learn more

Growth

Hiring, goals and reviews on the same records.

Goals & OKRs · Review cycles · Training · Surveys · Recruitment

Learn more

Comms

The messages that belong to the work, beside the work.

Chat · Mail · Helpdesk · Announcements · Notifications

Learn more

Business

Clients, deals and invoices, for the companies that also sell.

CRM · Invoicing · Reports · Import & export · Operator console

Learn more
05
Everything included

0 modules, one workspace, one set of identities.

Not a roadmap — built and running. The point isn't the count: the CRM deal, the person staffed on it and the invoice it produced are one join apart, not four exports apart.

Read how each domain works →

Employee records

PEOPLE

Profiles, departments, designations, prior experience, per-field change log.

NOT BUILT YET — SAYING SO NOW RATHER THAN DURING ONBOARDING

Native mobile apps, vendor-specific biometric drivers, a desktop activity tracker, an asset register, a Bikram Sambat calendar layer, leave accrual schedules and encashment, and shifts that cross midnight. The registered-device ingest endpoint and the data model behind biometric attendance exist; the device-vendor drivers do not.

06
What counts as late

Late is a line somebody drew. Move it.

One office, one month, every arrival plotted against the start of the shift. The grace window decides who was late; a second rule decides whether being late costs anything. Both are yours, and the second one ships switched off.

GRACE WINDOW

15 min

ARRIVALS COUNTED LATE

91

PEOPLE IT TOUCHES

56 of 100

DAYS OF PAY DOCKED

none

shift startson timelate from here-20-100+10+20+30+40

Drag the grace window

WHAT LATENESS COSTS

The grace window belongs to the shift, so a factory shift and an office one can disagree about it. The deduction is a company-wide switch that stays off until somebody turns it on, and it charges whole days rounded down — never a fraction of a day nobody agreed to. The arrivals above are a generated sample, not a customer’s month; the two rules acting on them are the ones the engine runs.

07
Payroll you can check

Try changing a salary rule.

Live, and it calculates the way your payroll would. Move anything — every figure shows how it was reached.

GROSS — monthly, the input80,000

Drag to change the salary

BASIC

Basic

percentage_of GROSS
%
48,000
ALLOW

Allowance

formula
32,000
PF

Provident fund

percentage_of BASIC
%
−4,800
SSF

Social security

percentage_of BASIC
%
−5,280
TAX

Income tax

slab_based
reads the slab table →
−3,242

annualised ÷ 12

NET PAYearnings 80,000 − deductions 13,322
66,678
take-home 83.3%withheld 16.7%
TAX SLABS — annual

Example rows. In a workspace these are records with a fiscal year on them.

0 – 500,000
%
500,000 – 1,000,000
%
1,000,000 – 2,000,000
%
2,000,000 – ∞
%
annual taxable  839,040

Why this matters

When a budget changes the rates mid-year, most systems make that a support ticket and a release you wait for. Here it is four numbers, edited by the person who read the circular, dated to the fiscal year it applies to — and the previous year's slabs stay exactly where they were, so last year's payslips still reproduce.

figures are illustrative · the real run evaluates server-side in a sandbox and writes an auditable payslip
08
Tax, drawn

The bracket you are in is not the rate you pay.

Move the salary. The staircase is the band your next rupee falls into; the curve is what you actually pay across the whole year. They are never the same number, and the gap between them is the thing people argue about at appraisal time.

ANNUAL SALARY

Rs 9,00,000

YOU ACTUALLY PAY

7.2%

YOUR TOP BAND

20%

0%10%20%30%marginal rate — your top bandeffective rate — what you pay5L7L10L

Drag the salary

These are the bands the payroll engine applies, and they are records in your workspace — one row per band, per fiscal year, editable by you. Every plan ships them unverified on purpose, with a screen that shows which rows nobody has checked against the current Finance Act yet. Rules you cannot inspect are rules you are trusting.

Run one month in parallel and check every figure against your spreadsheet.

That is the honest way to buy payroll software, and it is the step we would want if we were you. Nothing is finalised until you say so.

Start free — 15 daysBook a walkthrough

0

capabilities

across six domains, one workspace

0

permissions

short enough that somebody audits it

0+

backend tests

run before anything reaches you

0

database schema per company

not a tenant column somebody can forget
09
How it works

Claims are cheap. Here is how each one is actually built.

If a claim on this page has no mechanism behind it, it should not be on this page.

Nothing is approved by a message. An approval chain is an ordered set of steps; each decision writes an immutable row naming the actor, the verdict, the moment and the comment.

ApprovalChain
ApprovalStep
ApprovalAction
LifecycleEvent
10
Your data stays yours

Your data isn't filtered out of everyone else's. It was never in there.

Most multi-tenant systems put every company in the same tables and add acompany_idto every query. It works until one query forgets. Here each workspace is its own PostgreSQL schema, resolved before a query runs. Try both.

psql
SELECT * FROM payroll_payslip
  WHERE company_id = 42;
result setnot run
run the query to see what comes back
ISOLATION

One PostgreSQL schema per workspace, created and migrated at provisioning.

SESSIONS

Tokens held in HttpOnly cookies and proxied server-side — never readable by page scripts.

AUDIT

Employee changes, attendance edits, approvals and platform actions each write their own log.

EXIT

Your schema dumps cleanly. Portability is a property of the design, not a favour.

11
Built for your team

Where you are decides what you should turn on.

A five-person team switching on performance reviews is how software gets abandoned in month two. Pick yourself below, including what to leave alone.

5–30 PEOPLE
WHAT HURTS TODAY

HR is a founder's side task. Everything works until the first person asks how much leave they have left.

SWITCH ON FIRST

Employees, leave, attendance, self-service.

CAN WAIT

Payroll once headcount makes the sheet risky; recruitment when hiring becomes constant.

12
Getting started

Five steps, and one of them is us telling you not to trust us yet.

Migrations fail when a vendor asks for faith at the switchover. Step four exists because the only thing that earns trust in payroll is a month you checked yourself.

01

Claim a subdomain

Minutes, unattended

Your workspace is provisioned with its own database schema and its own migrations. Nothing about it is a row in a table shared with another company.

02

Bring your people in

One CSV

Import the spreadsheet you already maintain. Departments, designations and reporting lines come across; what doesn't map is reported back line by line rather than silently dropped.

03

Describe your rules

The real work

Salary components, tax slabs, leave types, approval chains, shifts, holidays. This is the part most systems hide behind a consultant — here it is a form you fill in and can change later.

04

Run a month in parallel

Our favourite step

Process a payroll period alongside your existing sheet and compare, line by line, before anybody depends on it. Trust the output because you checked it, not because we asked you to.

05

Operate

Month two onward

Payslips, reports, approvals and the audit trail behind all three. The spreadsheet becomes a backup you stop opening.

WHERE TO START

Pick the one that hurts most.

Nothing has to go live at once. Most companies start with the month they most dread and switch the rest on afterwards.

People

One record per person, and everything else reads from it.

The same employee exists three times — in a spreadsheet, in the payroll file, and in whoever knows they transferred in March. When the three disagree, payroll is wrong and nobody can say when it started.

7 capabilities

Discover

Time

Attendance, leave and shifts that agree with each other.

Leave is approved in email, attendance comes off a device, and the two only meet when payroll runs and somebody is docked for a day they were granted.

7 capabilities

Discover

Money

Payroll you can show your working for.

Payroll is a spreadsheet one person understands. When somebody asks why their tax changed, the answer takes an afternoon and rests on trusting that no formula was dragged one row too far.

9 capabilities

Discover

09
Answers

Ask the awkward ones.

Including the ones with a "no" in the answer. A vendor that only publishes the flattering questions is telling you which ones to ask on the call.

Each workspace gets its own PostgreSQL schema, created and migrated when the workspace is provisioned. Requests resolve to a schema before any query runs, so another tenant's tables are not merely filtered out — they are not on the connection's search path at all. There is no `company_id` column to forget in a WHERE clause.

10 of 10 answers shown
10
Talk to us

A person reads this.

Not a chatbot and not a drip sequence. Tell us what you're trying to fix and you'll get a straight answer — including "this isn't the right tool for you yet", when that's the honest one.

EMAILsales@deerx.com
ALREADY A CUSTOMERSign in at your own workspace address
WHAT BRINGS YOU HERE
NAME
EMAIL
ORGANIZATIONoptional
PHONEoptional
HOW MANY PEOPLE
WHAT ARE YOU TRYING TO FIXoptional

Stored as a lead record in the platform console. No newsletter, no third-party trackers, no reselling.

The spreadsheet was never the problem. Not being able to prove what it said last March was.

DeerX HRMS

A multi-tenant HR platform where the rules are records, the boundary is a schema, and the audit trail was there before anyone asked for it.

ON THIS PAGE
What it doesWhere the month goesSee the productThe six domainsEverything includedWhat counts as latePayroll you can checkTax, drawnHow it worksYour data stays yoursBuilt for your teamGetting startedPricingQuestionsTalk to us
ELSEWHERE
Create a workspacesales@deerx.com
© 2026 HRMS Platformbuilt in Nepal · deployable anywhere