Teams that run ERPNext quickly outgrow the built-in report views, Query Reports, and the dashboard charts on the Frappe desk. Analysts want ERPNext data inside the BI tools they already standardize on — Tableau and Looker Studio — where they can blend it with marketing, finance, and web data and share governed dashboards. The native path forces them into fragile CSV exports, hand-maintained ODBC connections straight at MariaDB, or one-off Python scripts that ignore ERPNext's permission model. None of those scale, and hitting the database directly bypasses role permissions, user permissions, and the multi-company boundaries ERPNext enforces at the application layer.

ECOSIRE builds a proper Frappe app — its own module, DocTypes, and hooks.py wiring — that publishes ERPNext data to Tableau and Looker Studio through governed, read-only endpoints. On the Tableau side we ship a Web Data Connector (WDC) that authenticates against ERPNext and pulls from curated extract definitions rather than raw tables. On the Looker Studio side we deliver a community connector (Apps Script) whose getSchema and getData map to the same whitelisted ERPNext methods. Every extract is a named Report Extract DocType you configure in the desk: which source (a DocType, a Query Report, or a curated SQL view we author), which fields, which filters, and which role profiles may read it. The connector calls @frappe.whitelist() methods over the Frappe REST API, so frappe.has_permission and user permissions are honored on every request — Tableau and Looker Studio only ever see rows the connecting service account is actually entitled to.

Refresh is incremental by design. Each extract declares a watermark field (typically modified, or a posting/transaction date), and the connector requests only rows changed since the last successful pull, so a nightly Tableau extract refresh or a Looker Studio cache warm doesn't re-scan millions of rows. A scheduler_events hook maintains lightweight sync state and can pre-materialize heavier extracts into a reporting table on a cron cadence, while doc_events in hooks.py (on_update/on_submit) keep near-real-time extracts current for the DocTypes you flag. Pagination, rate limiting, and typed field metadata (so Tableau sees dates as dates and currency as numbers) are handled in the connector layer, not left to the analyst.

Because this is build-to-order, nothing is pre-listed for instant download. You start with a short scoping call; we confirm the DocTypes, reports, refresh windows, permission mapping, and your Frappe/ERPNext version (v15 or v16 supported). We then build against a staging bench, validate through UAT, and install on your instance. Typical delivery is one working week from confirmed scope, depending on how many extracts and how much curated modeling you need. You receive the installable app source, the Looker Studio connector, documentation, and a support window — and the git repository is handed over so you own what we build.