Most biometric integrations for Odoo work by polling — a scheduled job repeatedly asks each ZKTeco, eSSL or Anviz device "any new punches?" over the SDK/TCP. That approach breaks the moment your Odoo server sits behind a firewall or NAT and can't open an outbound connection to every device, and it leaves attendance minutes (or hours) stale between poll cycles.

ADMS Push Attendance flips the direction. ZKTeco/eSSL/Anviz terminals natively support the ADMS (Automatic Data Master Server), also called the iClock/push protocol — the same mechanism ZKTeco's own ZKBioTime server uses. The device initiates HTTP(S) requests outbound to a listener endpoint, streaming each fingerprint, face or card punch the instant it happens. ECOSIRE builds a hardened ADMS listener into your Odoo instance so punches land in hr.attendance in real time, with zero polling and zero inbound firewall rules on the Odoo side.

This is a build-to-order module. It is not an existing apps.odoo.com download — ECOSIRE develops it against your exact Odoo version (17, 18 or 19), your device fleet and your network topology, then installs, configures and supports it. You receive the full module source, the device-side configuration, documentation and a support window.

How it works under the hood. The module ships a controller (http.Controller with auth='none' on the ADMS routes, protected by device serial + shared key) that speaks the iClock handshake: the standard GET /iclock/cdata registration and heartbeat, POST /iclock/cdata attendance-log payloads (ATTLOG, OPERLOG, user/fingerprint templates), and the GET /iclock/getrequest command queue for pushing commands back to the terminal. Incoming punch rows are parsed, de-duplicated and mapped to employees, then written into standard Odoo attendance models — so everything downstream (timesheets, Payroll working schedules, HR reporting, QWeb attendance reports) just works.

Data model & mapping. A new adms.device model (models.Model) stores each terminal's serial number, shared key, location and last-seen heartbeat, with a computed @api.depends-driven online/offline status field. A res.users/hr.employee device-ID map links each device user PIN to the correct employee — auto-created on first unknown punch and queued for review, or matched against a badge/PIN field you nominate. Raw payloads are archived in an adms.raw.log model so nothing is ever silently dropped and every punch is auditable.

Security is first-class. Access is governed by ir.model.access.csv plus record rules so HR sees attendance while the raw ADMS logs stay restricted to administrators. Device endpoints authenticate on serial + shared secret (and optionally an IP allowlist / mutual-TLS behind your reverse proxy), never on an Odoo session. Because devices talk out to Odoo over 443, you expose one HTTPS endpoint instead of poking holes to reach every terminal.

Automation & extensibility. Odoo automated actions / server actions can fire on each punch — flag missing check-outs, notify a supervisor on a late arrival, or trigger a Payroll recompute. Everything the module writes is reachable over Odoo's XML-RPC / JSON-RPC API, so external dashboards or a parent HRMS can read attendance without touching the devices. It runs on Community or Enterprise; Enterprise adds tighter Payroll/Planning integration where you have it.

ECOSIRE delivers the whole thing production-ready: listener, device provisioning, employee mapping, monitoring and handover training — typically in one working week depending on fleet size and network review.