LMCP for enterprise

Give teams useful AI without losing control of access.

Give your team governed AI access to Mail, Calendar, Teams, OneDrive, files, and business apps on Mac and Windows. Local integrations stay on the device; connected services use only the access your organization approves.

Mac + Windows
One deployment model
Local interfaces
Used where available
Folder scopes
Available for OneDrive
Explicit status
Shipped vs. roadmap

The problem

Why AI tools stall in enterprise

Most AI productivity tools require cloud API access, OAuth approvals, and token management. That means months of IT backlog before your team gets value.

API approvals take months

Graph API, Gmail OAuth, Exchange connector — each one requires a security review, a DPA, and sign-off from 3 teams. Your team waits.

☁️

Cloud tools create data residency risk

Sending emails and calendar events to a third-party cloud server triggers GDPR Article 28 obligations, cross-border transfer rules, and DLP exceptions.

🔑

Token management doesn't scale

OAuth tokens expire, scopes drift, and leaked credentials become incidents. Managing API keys across 50 machines is a full-time job.

Enterprise controls

Enterprise controls, built for IT

Core controls are live today. Governance features like per-tool disable and audit log are part of the pilot roadmap.

Available now
📁

OneDrive folder scoping

Restrict OneDrive access to specific folders per user or team. Agents see only what each role needs — no blanket access to the entire drive.

🔒

Platform-appropriate access

LMCP uses local app interfaces on Mac and Windows whenever possible. Integrations that require authorization are scoped, auditable, and limited to their intended service.

Pilot roadmap
🏢

M365 directory search

Query your enterprise Global Address List (GAL) via Microsoft Graph. Resolve first-name-only references in meeting notes to full names, titles, departments, and org hierarchy — no IT admin approval required for personal-scope queries.

🔧

Per-tool disable

Block any tool organization-wide with a single config flag. Restrict agents to read-only mail, disable file writes, or turn off Teams messaging — without touching each machine.

👁️

Read-only mode

Lock the entire installation to read-only. AI agents can query Mail, Calendar, Contacts, and Teams — but cannot send, create, or modify anything.

📋

Audit log

Every tool call is logged locally with timestamp, tool name, and input summary. Readable by your SIEM or compliance team via structured JSON.

⚙️

Config profiles

Deploy named profiles for different roles — Engineering (full access), Legal (read-only), Support (mail + calendar only). Manage them with your Mac or Windows device-management system.

📁

Global agencies

Enterprise teams

Used by teams at global agencies to connect AI assistants to OneDrive, Mail, and Calendar — with locally scoped access and reviewed connected services.

💡

OneDrive folder scoping in practice

Set onedrive_scopes in your config to restrict agents to specific folders only. Agents can read, list, and search within scope — and cannot access anything outside it.

Compliance

Governed access on
Mac and Windows.

LMCP keeps local integrations on each device and limits connected services to the scopes your organization approves. The exact data path depends on the app and platform, so security teams can assess each integration without relying on blanket claims.

Assess provider agreementsCloud AI providers and connectors may process request content. Review applicable DPAs and service terms.
Review retentionLocal data and logs remain under device controls; optional providers have their own retention policies.
Review transfersLocal tools run on-device. Cloud models, remote clients and connected services can transfer request content across networks.
Plan erasureUninstalling stops local LMCP processing; connected providers and stored records require separate review.
Works with existing DLPRuns inside your existing Mac and Windows management perimeter.
Assess the complete environmentLocal architecture supports your controls; it is not a SOC 2 certification or a guarantee of compliance.

Architecture comparison

Cloud-connected AI tools

Your MacCloud APIAI Model

Your data leaves the device. DPA required. Transfer risk.

LMCP

Your MacLMCP (local)AI Model

Data read locally. Your selected AI model receives requested content; review its hosting, retention and connected services.

Assess your data, providers and organizational requirements

Deployment

Live on your fleet in under 2 hours

Signed installers and managed configuration for Mac and Windows. Use Jamf Pro, Kandji, Microsoft Intune, or your existing device-management platform.

01

Choose the signed installer

Use the signed macOS package or Windows installer from local-mcp.com/download, then add it to Jamf, Kandji, Microsoft Intune, or your existing deployment system.

02

Push your config profile

Deploy the managed LMCP configuration through your Mac or Windows device-management system. Set folder scopes and policy keys for each role.

03

Apply platform permissions

On Mac, apply the required privacy preferences through MDM. On Windows, use your device and application policies for only the integrations your team enables.

04

Configure your AI client

Connect Claude Desktop, Cursor, VS Code, or another supported client. Each Mac or Windows PC uses its own LMCP instance and approved resources.

Config templates

Deploy now, grow with pilot controls

OneDrive scoping is live. Pilot customers can validate the broader governance roadmap before rollout.

Today — OneDrive scopingAvailable now
{
  "onedrive_scopes": [
    "Work Documents",
    "Shared/Team Projects"
  ]
}
Full EnterprisePilot roadmap
{
  "read_only_mode": false,
  "audit_log": true,
  "disabled_tools": [],
  "onedrive_scopes": ["Work Documents",
    "Shared/Team Projects"]
}
GDPR Strict / RegulatedPilot roadmap
{
  "read_only_mode": true,
  "audit_log": true,
  "disabled_tools": ["send_email",
    "onedrive_delete_file"],
  "onedrive_scopes": []
}
🚀

Start your 30-machine pilot

We’ll send you a signed pkg, a Jamf/Kandji deployment guide, and a GDPR review packet your DPO can sign off on. Most pilots are live within 48 hours.

✓ No commitment✓ Full enterprise controls✓ GDPR packet included✓ Jamf & Kandji guides