Ship the portal, the mobile app, the AI assistant, the automation — in days, not months. One gateway for every connector and every technology.
One gateway, four ways in. Read, write, run procedures, query across forms — over plain HTTPS.
Running procedures, reports, form actions, and smart filters is limited to the Priority WebSDK. Python, .NET, and mobile stacks can't reach those surfaces directly.
Connecting Lovable, Base44, or any LLM to Priority requires an MCP layer — and Priority doesn't ship one.
One Gateway opens both — and gives Priority the AI surface it never had.
One gateway sits between your customer's apps and their Priority. No SDK install on their side, no driver per laptop, no language lock-in — just HTTPS.
REST for code. SQLI for questions that cross forms. MCP for AI. One sign-in for the people using them.
The customer's portal, mobile app, n8n flow, Python script, .NET service — all hit the same REST endpoint. Forms, procedures, reports, document activations, file uploads. WebSDK and OData both sit behind it, and the gateway picks between them per call. One URL covers it.
POST /api/v1/query/form { "form": "ORDERS", "filter": { "STATDES": "Open" } }
Real SQL over Priority's own query engine: joins between tables, GROUP BY, aggregates — over plain HTTPS, with no driver installed on any machine. Read-only by construction: mutating verbs are rejected before the call leaves the gateway. Dates, shifted decimals, and Hebrew right-to-left text come back decoded, not as raw Priority integers.
POST /api/v1/sqli/query SELECT ORDERS.ORDNAME, CUSTOMERS.CUSTDES FROM ORDERS, CUSTOMERS WHERE ORDERS.CUST = CUSTOMERS.CUST FORMAT;
Plug Claude.ai or Claude Desktop into the gateway and the customer's IT manager asks live questions of their Priority. Or build the customer's own AI assistant on top of the same MCP surface.
▸ You Show open orders for ACME. Claude priority_action(query, ORDERS, { CUSTNAME: "ACME" })
Server-side callers post credentials once to /auth/login and get back a bearer token with an idle window, an absolute expiry, and a per-user cap. Browser apps embed the hosted Connect popup instead: the user signs in to Priority inside it, and the app receives a revocable token. The password never reaches the calling page.
POST /api/v1/auth/login → { "token": "gws_…", "expiresAt": "2026-08-08T18:20:00Z" }
Added since the gateway's first release.
The gateway builds a scoped schema context out of the customer's own Priority — the tables a question touches, the filters that give them meaning, the join paths between them — and hands it to the model before it writes a line of SQL. "How much did we sell last month" comes back right the first time. Off by default; needs the BI schema graph on the same host.
The Priority WebSDK allows one login per process, so simultaneous calls used to queue behind one another. Turn on the worker pool and the gateway spreads them across a pool of worker processes, each holding its own Priority login.
/api/v1/sqli/syntax returns the SQLI dialect. /api/v1/guide answers which technology fits a given job. /api/v1/n8n-guide is the n8n node's own authoring spec. A model fetches what it needs mid-conversation instead of guessing.
Shopify orders into Priority. Nightly invoice exports out. Document approvals routed through Slack. Drop the n8n-nodes-priority-gateway node onto the canvas and connect it — no bespoke glue code, no per-customer SDK build.
/api/v1/n8n-guide, so an LLM can write a working workflow without opening the editor.v3.17.1 · 7,657 npm installs in the last 12 months · 61 releases
Intake, live metadata discovery, scaffolding, a real-browser smoke test, and a handoff zip — one workflow. You describe what the customer needs; the generator ships a runnable Priority-backed app.
Three questions lock pattern, form, and connection mode.
Researcher agent pulls live form metadata from the gateway.
Builder agent scaffolds the app on disk under generated-apps/.
Tester agent walks the scenario in a real browser. Up to three retries.
Refinement re-enters at the lowest applicable phase — no full rebuilds.
Pick the shape, pick the stack — the generator picks the template.
| React | HTML | Python CLI | |
|---|---|---|---|
| Web app | forms & auth | single-file | — |
| Dashboard | KPIs & filters | read-only | — |
| Mobile-first | responsive | PWA, offline-resilient | — |
| Automation | — | — | cron-friendly |
The orchestrator dispatches one of these per phase.
| Agent | Responsibility |
|---|---|
| Researcher | Pulls live form metadata from the Gateway |
| Designer | Maps forms and subforms to UI patterns |
| Builder | Scaffolds files under generated-apps/ |
| Tester | Walks the scenario in real Chrome (≤3 retries) |
Reference apps the generator copies and customizes. New patterns ship in every release.
orders-dashboard-react
Read-only orders dashboard with KPIs, search, and status filter.
stock-count-mobile-html
Mobile PWA stock count with subform writes and a finalize activation.
invoice-export-python
Cron-friendly CLI that exports AINVOICES rows to CSV.
employee-order-react
Employee-facing order entry form backed by Priority sessions.
auth-contact-react
Contact form behind contact-based authentication.
auth-priority-user-react
Auth scaffold for Priority-user logins.
contacts-list-session-react
Stateful list browser with debounced search via the Session API.
+ more coming
New patterns ship in every release.
Your customer wants AI that knows their Priority. Plug Claude.ai or Claude Desktop into the Gateway and they get it — their forms, their fields, their data, in their own words.
The LLM learns your forms, fields, subforms, and dropdown options on demand.
➔"Show open orders for ACME from this month." Natural language in, structured rows out.
➔Create records, run procedures, generate reports, upload files — all in one conversation.
Works with
Chat with live Priority data — ask, query, create.
Generate web apps that read and write Priority through the gateway.
Internal dashboards on top of Priority — no backend code.
LLM reasoning inside workflows — smart routing, alerts, and triage.
Any MCP-compatible client connects. The gateway handles auth, routing, and session lifecycle. Full MCP details →
Real customer asks already running on the gateway.
Orders from e-commerce, EDI, or email flow through n8n into Priority — validated, enriched, and booked.
Single-file PWA on warehouse phones, generated from one prompt. Subform writes, finalize activation, survives a closed tab.
"Which orders shipped late this month?" Claude reads live Priority data and answers — no custom report.
Orders, invoices, and delivery status — built in Lovable or Base44, served by the gateway.
Python CLI on a cron pulls AINVOICES to CSV, drops it on S3, SFTP, or FileSmile. Generated from one sentence.
Account ledgers and statements triggered from external systems — full document-activation flow over HTTP.
Hosted for you on Amazon Web Services with Cloudflare in front, or installed on your own network as a single Windows service. Same gateway either way.
The hosted gateway runs in Amazon's Tel Aviv region, deployed from a versioned pipeline. Secrets live in AWS Parameter Store, encrypted, never in a database or a config file.
Every request arrives through Cloudflare: TLS, DDoS protection, a web application firewall, rate limiting and bot filtering, plus caching and routing that shorten the trip from wherever your callers are.
The servers accept no inbound connections. They dial out to Cloudflare and traffic comes back down that tunnel, so there is no public port to find and nothing to scan. Admin surfaces sit behind a second gate.
Running Priority on your own premises? The gateway ships as a single Windows service, and the same Cloudflare model is how it reaches the outside world — without opening a port on your network.
Part of the Priority AI-Tools suite
One registration connects your team to Priority through AI — Claude, ChatGPT, Gemini, BI, and app generation.