Jobs — a record, not an advert+
Ten tabs on one job: overview, applications, pipeline, recommended, interviews, publishing, notes, activities, documents and email. Five saved views you work from rather than filters to rebuild every morning — assigned to me, active, all, draft, completed. Creation runs in four gated steps: job information, JD generator, templates and commercials, then overview and posting. The seven-stage pipeline is applied, screening, interview, assessment, offer, hired and rejected, and dragging a card advances the applicant and stamps the trail. Professional and soft skills each carry a weighting plus the share of applicants holding them, and matching scores against it. Every job has its own inbound email address, so an application arriving by mail lands on the record.
Candidates — a database you can actually work+
Not a folder of CVs. The row itself carries candidate, top skills, experience, score, live applications and status, so a shortlist needs no clicks. Seven tabs per person: CV, applied jobs, interviews, offers, compliance, documents and activity. The original document is kept and parsed out as data, which is what makes someone findable. One person can be live on several roles at once with the stage and score they hold on each. Compliance is tracked as individually named requirements in packs, so “not cleared” always says which item is actually outstanding. And every change, stage move and message is dated and attributed — the next recruiter inherits the context, not just the CV.
Resume parser — six stages you can watch+
Ten files a batch, with the limit stated on screen and a counter for the room left. Six named stages: uploading, extracting text, detecting sections, extracting fields, normalising and scoring, then matching candidates. A live activity log writes out every step as it happens and is retained after the run. The stages are named for a reason — a file that fails points at the stage it failed on rather than reporting an error. Full name and email are required; phone, location and summary are not, so a partial parse is still usable. And nothing is auto-imported: results are reviewed first, and creating the candidate is a separate, deliberate action.
Candidate matching — the database ranks itself+
Selecting a job loads its own skills, experience band and location as the starting criteria, so a search begins from the brief. Five criteria — skills, job title, experience, education, location — each switching on independently and contributing nothing until it does, with a slider deciding how hard each one pulls. The score is printed on every row, so a low match can be judged on its merits rather than quietly hidden. The whole database is ranked, not a top five cut off at an arbitrary line, which is what makes the long tail usable. Rows are selectable for bulk action, and a criteria set can be saved and run again.
Offer letters — not a mail merge+
A searchable template library — standard offer, contract engagement, internship, executive — each carrying its own fields, with the candidate’s personal information, address, preferences and compensation sitting beside the document so a letter is filled, not retyped. Nineteen placeable field types in four groups: the ones that execute the letter (signature, initial, stamp, data sign), text elements (text, number, checkbox, radio, dropdown, date, table), personal data pulled from the record, and extend elements including image, attachment, approve and decline. Approve, decline and reset live on the letter itself, and the outcome lands on the candidate’s offer tab and the job’s pipeline.
Dashboard and settings — configuration, not code+
Four live counters against last month, the pipeline funnel with headcount and conversion at each of the seven stages, recent jobs, AI match alerts since you last looked, and a dated activity feed drawn from every module. Job status mix, source attribution and recruiter KPIs against target sit on the same page — analytics without a BI seat. Underneath, settings covers users and five roles with each role’s permissions listed rather than implied, job code format and pipeline stages, candidate settings including duplicate detection, consent and retention periods, KPI targets with monthly bands, and the template libraries and named approval routes reused by every record.
