AI integrations

Your AI can act on your operations — safely.

BoringOps exposes a native MCP server, so Claude — or any MCP-compatible AI — can trigger, approve, and manage operations directly, with full audit trail and governance.

Get early access → Lihat cara kerjanya

MCP

Native server

AI models supported

100%

Audit trailed

Claude · BoringOps MCP

Terhubung · tools operasional aktif

Approval apa yang masih pending untuk Pak Budi?

Budi has 3 items waiting for approval:

Restock Gudang APending

Arief · Warehouse

Overtime — WinaPending

HR · Finance

Vendor payment — PT MajuPending

Finance · Rp 12.5jt

Summarise overdue tasks this week by department

Overdue this week — 5 items

Warehouse2 overdue

CS Team2 overdue

Finance1 pending

Cara kerja

Connect once. Let your AI do the asking.

Add BoringOps as an MCP server in your AI client and you're done — no custom code, no webhooks to wire up.

Step 01

Add the MCP server

One JSON block in your Claude Desktop or AI client config. Takes 60 seconds.

Step 02

Authenticate with your API key

Set your BORINGOPS_API_KEY env variable. Scoped to your workspace.

Step 03

Ask in plain language

Your AI can list tasks, check approvals, query by department or status — in Bahasa or English.

Step 04

Every action is logged

AI-triggered approvals carry the same full audit trail as manual ones. Nothing struks through.

Setup

Add BoringOps to Claude Desktop in 60 seconds

Edit your claude_desktop_config.json to register the BoringOps MCP server. Once connected, Claude can query and manage your operations in plain language.

Works with Claude Desktop, Cursor, and any MCP-compatible client

API key scoped per workspace — revoke anytime from the dashboard

No data stored by AI providers — queries stay inside BoringOps

          
claude_desktop_config.jsonMCP config

        // Add BoringOps to Claude Desktop

        {

          "mcpServers": {

            "boringops": {

              "command": "npx",

              "args": ["@boringops/mcp-server"],

              "env": {

                "BORINGOPS_API_KEY": "your-key"

              }

            }

          }

        }

Model agnostic

Plug in the AI you already trust

BoringOps doesn't lock you to one model. The operational layer stays consistent regardless of which AI you run.

🤖

Claude by Anthropic

Excellent Bahasa Indonesia understanding. Strong on long-context documents and nuanced approval reasoning. Recommended for Indonesian SMEs.

Recommended

✨

GPT-4o by OpenAI

Strong general reasoning and multimodal document parsing. Good for teams that already use ChatGPT Enterprise or the OpenAI API.

🔷

Gemini by Google

Deeply integrated with Google Workspace. Ideal if your team runs heavily on Sheets, Drive, and Calendar for operations today.

🔌

Any MCP-compatible AI

BoringOps implements the open MCP standard. Any client that supports MCP can connect — today or in the future.

Capabilities

What your AI can do with BoringOps

Every action is permissioned, logged, and reversible — the AI can act, but you stay in control.

📋

Query pending work

"What's waiting for approval?" — get a structured list across all departments, filtered by assignee, date, or status.

⚡

Trigger approvals

Approve or escalate a request on behalf of an authorised user — with a mandatory reason captured for the audit log.

📊

Summarise by department

"Which teams are bottlenecked this week?" — AI generates a plain-language ops summary from live data.

🔍

Search the audit trail

"Who approved the Rp 50jt payment last month?" — trace any decision back to the actor and timestamp.

📢

Draft comms

Generate a WhatsApp or Slack summary of this week's overdue items — ready to send to the relevant team lead.

🛡️

Governed by permissions

Every MCP tool is scoped to your role. A finance AI agent can't see HR records — the data boundary is enforced server-side.