Book a demo
Arpuro Intelligence·Now in early access

Ask your factory anything.

Arpuro Intelligence turns four billion dust measurements into answers — in plain language, in real time. Connect Claude, ChatGPT, Cursor or any MCP-aware client and let your shift leads, HSE engineers and operators query live data without writing a single line of SQL.

Book a demoRead the technical docs
4B+
measurements indexed
20
industrial sites live
3
countries · NO · IS · ES
15%
of European Si smelting
The shift

Same question.
Four very different paths.

Three of these require an engineer, a dashboard and a deployment. One requires a sentence. That's the difference Arpuro Intelligence makes.

The question

"Show every location at Isfjorden Silicon with last month's dust conditions, sorted by the highest exposure first."

Approach 01SQL
SELECT l.name,
       AVG(m.resp_mg) AS avg_dust
FROM measurements m
JOIN locations l ON m.loc_id = l.id
JOIN sites s     ON l.site_id = s.id
WHERE s.name = 'Isfjorden Silicon'
  AND m.time >= NOW() - INTERVAL '1 month'
GROUP BY l.name
ORDER BY avg_dust DESC;
Needs: schema, access, SQL skills
Approach 02REST
GET /api/sites
  ?name=Isfjorden+Silicon

GET /api/locations
  ?site_id=42

GET /api/measurements
  ?locations=101,102,103,104
  &start=2026-01-01T00:00:00Z
  &end=2026-02-01T00:00:00Z
  &types=resp_mg_per_m3
Needs: API keys, glue code, sorting client-side
Approach 03Python
sites = api.get_sites(
    name="Isfjorden")
locs = api.get_locations(
    site_id=sites[0].id)
data = api.get_measurements(
    locations=[l.id for l in locs],
    start="2026-01-01",
    end="2026-02-01",
    types=["resp_mg_per_m3"])
sorted(data,
    key=lambda x: x.avg,
    reverse=True)
Needs: SDK, Python env, deployment
Approach 04MCP

"Show every location at Isfjorden Silicon with last month's dust conditions, sorted by the highest exposure first."

A sentence. The model does the rest.Arpuro
What you can ask

From shift handover to board reporting.

Tap any prompt to see what the answer looks like.

Operations

Show me the highest dust readings from Furnace 4 last shift.

Furnace 4 — last 8 hours (06:00–14:00 UTC)

Three respirable dust peaks crossed the 15-minute OEL of 5.0 mg/m³. The highest, at 11:42, hit 7.8 mg/m³ during a tap. Average across the shift sat at 1.4 mg/m³ — well under the 8-hour limit of 2.5.

Respirable dust · mg/m³ · 15-min buckets06:00 → 14:00
15-min OEL · 5.08h OEL · 2.5842006:0008:0010:0012:0014:00
TimePeak (resp)% of15-min OELLikely cause
07:305.3 mg/m³106%Charging
11:427.8 mg/m³156%Tapping
13:006.5 mg/m³130%Slag handling

Want me to break this down per 15-min window, or compare against last week's same shift?

Analysis

Compliance

Reporting

How it works

Built on an open standard.

Arpuro Intelligence speaks the Model Context Protocol — the open standard for letting LLMs talk to live data. Think of it as USB-C for AI: one connector, every modern model.

97M+
SDK downloads
per month
13k+
MCP servers
on GitHub
Backed by
Anthropic · OpenAI · Google · Microsoft
  1. 01

    Add the connector

    In your MCP-aware client (Claude, ChatGPT, Cursor, Copilot), paste a single URL:

    https://mcp.arpuro.io
  2. 02

    Sign in once

    A standard OAuth flow against your Arpuro account. The same login as the platform — same role, same permissions, no parallel access management.

  3. 03

    Ask anything

    The model gets read-only access to the dashboards and history you already have permission to see. Charts, summaries, exports — all generated from live data.

Integrations

Works with the tools your teams already use.

AI clients
  • ClaudeAnthropic
  • ChatGPTOpenAI
  • CursorIDE
  • CopilotMicrosoft
Workflow & automation
  • n8nAlerts & reports
  • OpenClawAgent orchestration
  • Zapier / MakeTriggers
  • Custom MCP clientsOpen standard
Enterprise
  • Power BIReporting
  • Microsoft TeamsAlerts
  • SAPDirect API
  • Word & ExcelSubmission-ready
Tangible RoI

Where the numbers land.

An independent assessment by Crisp Ideas in April 2026 modelled the per-site impact of combining Arpuro sensors with generative AI. The picture isn't a productivity tweak — it's RoI amplification on top of physical infrastructure.

Read the full assessment
Direct savings
0.7–3.4M
NOK per site, per year. Cleaning cycles, campaign replacements, regulatory risk avoidance.
Societal value
10.8M
NOK per typical case, DALY-adjusted (Disability-Adjusted Life Years).
CAPEX reframed
Tens of millions
NOK per ventilation install. Sensor data tells you whether you need it — before you spec it.
Operational wins
Clean less, fail less.
Periodic cleaning at NOK 100k per cycle, performed when needed — not on the calendar.
Security & access

Read-only by default. Scoped to what each user already has.

  • Tokens are short-lived and revocable from the Arpuro dashboard.

  • The model never sees data without an explicit user prompt.

  • Built on existing platform RBAC — no parallel access plane.

Open the technical docsTalk to an engineer

Hand your sensors a microphone.

Thirty-minute walkthrough on your data, with one of the engineers who built the platform.

Book a demoContact sales