Janus Gate Logo
Janus Gate
Local attendance via face matching
FREE • LOCAL • WINDOWS
No cloud. No theatre. Just a local time clock.

Attendance, on one PC

Enroll • Register • Query • Export

Janus Gate is a simple attendance system (time clock) using face matching, running locally on any Windows PC. It is free to use.

  • Enroll: name + section, capture a live photo (exactly one face).
  • Register: look at the camera and click REGISTER to clock IN/OUT.
  • Query: filter by date range, name, section, and record type; export CSV or print.
  • Charts: optional chart view when you need a quick overview.
Download ↗ How it works ↘
This does not “prove identity”. It matches similarity. Lighting, angle, glasses, ageing, and camera quality will affect errors.
UI: images/janus_register.png
Janus Gate Register UI Register

What it is

Local PC • Face matching • Attendance logs

A local time clock, by face matching

  • Runs locally on a Windows PC.
  • Keeps an internal database with facial “signatures”.
  • Stores timestamped IN/OUT records and work-session durations.
  • Free to use.

What the user actually does

  • Look at the camera + click REGISTER to clock in/out.
  • Open Query to view/export reports when needed.
Simple point: replace manual sign-in sheets with a quick local workflow.

Enroll

Name • Section • Live capture

Register a person

  • Enter Name and Section (example: “Warehouse A”).
  • Take a live photo using the camera.
  • The capture must include exactly one face.
  • If the capture has zero faces or multiple faces, it is rejected.
  • The system saves a facial “signature” in the internal database.

Practical note

  • Use decent lighting and a front-facing pose.
  • Bad input equals bad matching later.

Register

Recognize • Toggle IN/OUT • Durations

Clock IN/OUT with one button

  • Camera capture + face matching.
  • If not recognized: shows “Not recognized” and tells the person to go to the registration desk.
  • If recognized: automatic toggle.

Toggle logic

  • If there is no open entry: creates an IN.
  • If there is an open entry: creates an OUT and calculates session duration.
  • If the session is < 5 hours, it assumes it might be a break and also shows the combined duration of recent sessions.
Day-to-day usage is deliberately dull: look + click.

Query

Filters • Totals • CSV • Print • Charts

Search and filter

  • Go to the Query page.
  • Select a date range.
  • Filter by Name, Section, and record type:
    • All
    • Complete (IN + OUT)
    • Open (IN only)

Outputs

  • Results in a table with totals (record count, total hours, and related summaries).
  • Export CSV for payroll/HR workflows.
  • Print for paper trails.
  • If “Chart” is selected, it shows a chart view.

Limits & risks

False negatives • False positives • Biometric data

This is not identity proof

  • It is similarity matching, not verification of “who you really are”.
  • Lighting, camera angle, glasses, ageing, and camera quality affect results.
  • Bad threshold tuning gives:
    • false negatives (does not recognize)
    • false positives (worst case: recognizes the wrong person)

Local does not mean harmless

  • It runs locally, but it is still biometrics.
  • If someone gets access to the files/database, they get sensitive material.
  • Lock the PC down, restrict access, and treat backups as confidential.
If you deploy it carelessly, you are not “innovating”. You are just creating a new failure mode.