MCP server exposing Acumatica ERP (contract-based REST API) as tools for AI agents. Multi-instance, OAuth2.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"grp-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# grp-mcp MCP server that exposes **Acumatica ERP** as tools for AI agents. Multi-instance, OAuth2 — point it at any Acumatica site with a base URL + credentials. It reaches Acumatica through **five client planes** (four for driving, one diagnostic-only), so an agent can read and write almost anything: - **Contract-based REST** — CRUD entities, bulk-load from Excel/CSV, invoke actions, run reports, attach files, manage customization projects. - **DAC-based OData** — read tables/DACs that aren't on the endpoint, plus mandatory-field metadata (`run_dac_odata`, `get_dac_metadata`). - **Screen-based SOAP engine** — *drive screens the REST API can't*: context / master-detail / wizard screens (segment values, Enable Features, the financial calendar…). Discover (`list_screens`, `screen_get_schema`), read (`screen_get`), and write (`screen_submit`, with dry-run + per-field errors + dialog auto-answer) any screen by replaying its UI commands — no browser, no zeep. Higher-level writers (`screen_insert_rows` for master-detail/bulk grids, `screen_record` for idempotent create-or-edit) and ready-made setup recipes (`create_financial_calendar`, `create_ledger`, `chart_of_accounts`, `enable_features`, `manage_financial_periods`) sit on top. - **Modern UI-screen plane** — the JSON protocol the real browser UI uses (`/t/<Tenant>/ui/screen/<ScreenID>`), reusing the SAME login session as the classic plane (no extra auth, no browser). `ui_get_structure` reads a screen's full descriptor — views, fields, **enum allowed-values**, action inventory, grid keys — and `ui_screen_action` sets fields + fires actions (incl. dialog-confirm). This reaches what classic SOAP can't: actions whose classic tag is a silent no-op (e.g. GL201000 Generate Calendar), enum value discovery, live workflow-aware field/action state, and **full grid CRUD** — read, insert, update, and delete an *existing* grid row (the classic engine can only append rows; editing one in pla…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.