From careers page to signed offer, in one product.

Applications, AI screening, WhatsApp and voice outreach, scheduling, scorecards and offers. Built for bulk and lateral tech hiring in India.

  1. Careers page

    Roles published from the ATS. Applications land in the pipeline.

  2. AI screening

    Every resume scored against the role blueprint, with its evidence.

  3. Outreach

    WhatsApp templates and AI voice screening from one queue.

  4. Scheduling

    Panelist availability, candidate slot picking, calendar invites.

  5. Scorecards

    A structured scorecard per round, not a note in a thread.

  6. Offer

    Approval chain, e-signature, and the audit trail behind both.

One candidate, one record, six stages. No re-keying between tools.

Measured on our own hiring, since July 2026. NxtWave runs its engineering and instructor hiring on this product.

Applications processed
6,100+Applications processed
Interviews scheduled
220+Interviews scheduled
AI screening calls placed
3,500+AI screening calls placed

Everything between an application and a hire

Most teams run this on six tools that do not talk to each other, and lose candidates in the gaps between them. This is the same work in one record.

Resume screening that shows its work

Every application is scored against the role's blueprint, and the score arrives with the lines of the resume that produced it. A recruiter can disagree with it in one click, and the override is recorded.

WhatsApp and AI voice, from one queue

Template messages, replies routed back to the candidate record, and AI screening calls placed from the same list your callers work. Suppression and do-not-call are enforced at the point of dialling, not per campaign.

Scheduling without the thread

Panelist availability, candidate slot picking, calendar invites and reschedules.

Structured scorecards

A scorecard per round against the same rubric, so two panelists are answering the same question.

Offers with an approval chain

Approvals, e-signature and an audit trail of who changed what, and when.

What the first two weeks look like

  1. Bring your roles

    Import open requisitions and the rubric each one is judged against, or start from a template. Your careers page publishes from the same record, so a role goes live once.

  2. Connect the channels

    Your WhatsApp sender, your mail domain, your panelists' calendars. Candidates begin arriving in the pipeline instead of in an inbox.

  3. Run a real requisition

    Screen, call, schedule, score and offer against one live role, end to end, before you move the rest of the team across.

It plugs into the tools already running your hiring

  • WhatsApp

    Template sends and inbound replies, routed to the candidate record.

  • Gmail and Amazon SES

    Transactional and campaign mail from your own verified domain.

  • Google Calendar

    Panelist availability and interview invites, with reschedules.

  • Darwinbox

    Requisitions and joiner data synced with your HRMS.

  • Zoho Sign

    Offer letters out for e-signature and back, tracked in the pipeline.

  • Stripe

    Subscription billing and metered usage for your workspace.

What we will put in writing about your data

You are handing us candidates' names, phone numbers and resumes. These are the three things worth checking before you do.

Separation is enforced on every query, not by convention

Every row carries the workspace it belongs to, and the storage layer appends that predicate to each query it builds rather than trusting the calling code to remember. Postgres row-level security policies keyed on the same workspace claim sit behind that as a second layer, and are being switched on workspace by workspace.

Your data stays in India

The database runs in AWS Mumbai and the application runs in Vercel's Mumbai region. Candidate documents live in object storage in the same region.

Consent, export and deletion are part of the product

A candidate's consent is recorded with its timestamp and source when they apply. You can export your whole workspace on demand. Deleting a workspace happens in two steps thirty days apart, so a mistake is recoverable, and the rows are then purged rather than hidden.

Read the full security note for the detail your security review will ask for.

Questions we get asked first

Is this only for engineering roles?

It is built for technology hiring, which is where the screening rubrics and the call scripts are strongest. Nothing in it is restricted to engineering, but a non-technical requisition gets less out of the screening than a technical one does.

Does it replace our HRMS?

No. It runs the hiring up to the signed offer. Darwinbox and systems like it keep the employee record, and requisitions and joiner data sync between the two.

What does it cost?

Plans and their limits are on the pricing page. We have not set final prices, so the page tells you what each plan includes and asks you to talk to us for the number rather than quoting one we would have to change.

Can a candidate see what the AI decided about them?

A score never rejects anyone on its own. It arrives with the evidence behind it, a recruiter can override it in one click, and the override is recorded against the application.

How much of our process do we have to change?

The stages, the rubrics, the rejection reasons and the mail and message templates are yours to configure. The parts that are not configurable are the ones that exist to keep an audit trail honest.

What do we need before we start?

Your open roles and what each one is judged on. A WhatsApp sender, a mail domain and your panelists' calendars can be connected afterwards, in whatever order suits you.

Run your next requisition on one system

Start on the free plan with one open role, or bring us a requisition you are struggling with and we will walk through it together.