Skip to content
MCP-native · Desktop · Web · Code
PortEden

Secure Your Data With Claude

PortEden makes Claude better at your real work. Connect via MCP in Claude Desktop, claude.ai, or Claude Code, and Claude picks up redaction, per-tool RBAC, and an audit log on every tool call — so the same Claude your team uses every day operates on real client work without leaking the underlying data to the model.

See pricing

Free for solo · No credit card · 5-min MCP setup

Works With What You Already Use
Gmail
Gmail
Outlook
Outlook
Google Calendar
Google Calendar
Google Drive
Google Drive
Google Docs
Google Docs
OneDrive
OneDrive
SharePoint
SharePoint
Teams
Teams
Slack
Slack
Notion
Notion
Asana
Asana
Monday
Monday
Linear
Linear
Jira
Jira
Confluence
Confluence
Entra ID
Entra ID
Better Together

What PortEden Adds to Claude

  • Claude reasons over the context it receives. PortEden enforces redaction at the firewall layer before tool-call responses reach Anthropic — sensitive identifiers (names, contact details, account numbers, free-text identifiers) are tokenised in transit, so the model produces useful output while the underlying values stay inside your tenancy.
  • Claude inherits your identity model. PortEden synchronises with your IdP (Okta, Entra, Google Workspace), so user permissions and group membership flow into the firewall automatically — when someone leaves the team or changes role, their Claude tool access updates in the same provisioning cycle as the rest of their accounts.
  • Claude Desktop, Claude Code, and claude.ai all speak MCP. PortEden registers as a first-class MCP server on each surface, applying the same role-based access control, policy groups, and per-contact rules across every Claude entry point your team uses.
  • MCP gives Claude tool use; PortEden adds the governance around it. Every tool call is recorded with the request, decision, and redacted response in an exportable audit log; access decisions follow your identity-sync state and policy groups; revocation propagates across Desktop, Web, and Code immediately.
Anthropic Privacy Policy
Capability Matrix

Claude With PortEden: What You Get

Real-time PII redaction in tool responses
50+ identifier types stripped before data reaches the model
ClaudeNo
+ PortEdenYes
Per-tool, per-action RBAC (read vs. write vs. delete)
ClaudeNo
+ PortEdenYes
Per-contact / per-domain firewall rules
ClaudeNo
+ PortEdenYes
SIEM-exportable audit log of every tool call
Anthropic Console covers admin events; PortEden adds per-tool-call detail
ClaudePartial
+ PortEdenYes
Native Model Context Protocol (MCP)
Anthropic created MCP; PortEden ships 6 MCP endpoints
ClaudeYes
+ PortEdenYes
Token-context reduction (smaller, faster prompts)
~80% reduction via context hygiene
ClaudeNo
+ PortEdenYes
IdP identity sync (Okta / Entra / Google Workspace)
Permissions revoke with the rest of the user's accounts
ClaudeNo
+ PortEdenYes
Confirm-before-write on destructive actions
Send / delete / forward require explicit confirmation
ClaudeNo
+ PortEdenYes
One-click revocation across all connected tools
ClaudeNo
+ PortEdenYes
MCP / Agentic Coverage

Every Claude Surface, One Firewall

Connection: Native MCP — 6 endpoints, 50+ tools

Claude Desktop (macOS / Windows / Linux)

Add PortEden as an MCP server in claude_desktop_config.json via mcp-remote. PortEden enforces per-tool permissions before Claude sees anything.

Setup guide

claude.ai (Web)

Add PortEden under Customize > Connectors. OAuth flow takes 30 seconds; tokens auto-sync to Claude Desktop.

Setup guide

Claude Code (CLI / IDE)

Same MCP endpoints, stdio bridge. Per-token quotas keep agentic loops bounded; PortEden audits every tool call.

Setup guide
Tooling Notes
  • 6 PortEden MCP endpoints: email, calendar, drive, google-docs, google-sheets, tasks — 50+ individual tools.
  • Tool-use, function-calling, and streaming all pass through the firewall — same redaction and RBAC for ad-hoc chats and long-running agents.
  • Per-token permissions inherited from my.porteden.com: scope by provider, account, time window, and field visibility.
  • Per-contact and per-domain access rules — block individual senders or attendees from any AI request.
  • Monthly request quotas per token; revoke instantly with one click across all Claude surfaces.

Connect Claude in 5 Minutes

1

Sign in to PortEden and connect a data source

OAuth into Gmail, Outlook, Calendar, Drive, or Tasks at my.porteden.com. Pick the access profile that matches your industry.

2

Add PortEden's MCP endpoints to Claude

On claude.ai use Customize > Connectors. On Claude Desktop, paste the mcp-remote config. Six endpoints, one auth flow each.

3

Use Claude exactly the way you do today

Ask Claude about anything in your stack. PortEden redacts, audits, and enforces RBAC on every request — invisibly.

Read the full Claude Desktop guide →
Claude + PortEden

Five-Minute Setup. Free While You Test.

Connect a data source, plug Claude into PortEden, and put Claude to work on the data your team actually needs to handle.

Read the Claude guide

Frequently Asked Questions

Does Claude actually support MCP, or is this a custom integration?
MCP is native. Anthropic published the Model Context Protocol spec; Claude Desktop, claude.ai, and Claude Code all consume MCP servers as a first-class feature. PortEden ships six MCP endpoints — Claude calls them the same way it calls any other MCP server, and the firewall is invisible to the model.
What's the difference between connecting via Claude Desktop and via claude.ai?
claude.ai uses the Connectors UI — a 30-second OAuth flow per endpoint, no config file. Claude Desktop uses claude_desktop_config.json with mcp-remote to bridge remote MCP servers into stdio. Connectors added on claude.ai automatically sync to Claude Desktop, so most users start there. Claude Code uses the same MCP endpoints via the CLI.
What happens to a Claude tool call when PortEden blocks something?
The MCP tool returns a structured permission-denied response. Claude sees the rejection, explains it to the user ("I tried to send to that contact but your firewall blocked it"), and the audit log captures the attempt — including the tool, the arguments, the rule that fired, and the timestamp. No silent failures, no model retries against a different tool.
Will PortEden slow Claude down?
Redaction adds ~150-200 ms per tool call. Because PortEden also reduces token usage (~80% less context per request), most users see end-to-end latency drop, not increase. Claude returns answers faster on PortEden's clean payloads than on the same provider's raw API output.
Does this work with Claude Code and agentic loops?
Yes. Claude Code consumes the same MCP endpoints over stdio. Per-token monthly quotas keep agentic loops bounded; per-tool RBAC blocks destructive actions (delete, send) when you don't want them; the audit log captures every tool call in the loop, exportable as CSV or piped to a SIEM.
How do I revoke Claude's access?
Sign in to my.porteden.com and revoke the token. The change is immediate across Claude Desktop, claude.ai, and Claude Code — Anthropic's next tool call gets a 401, and Claude prompts for re-auth. No need to delete connectors in Claude itself.
Is there a free tier?
Yes — free for solo licensed practitioners and small developer use. Higher monthly tool-call quotas, SSO, and SIEM export are on paid plans. See pricing for details.

Keep Exploring

Get More From Claude With PortEden

Five-minute setup. Free tier for solo licensed practitioners. Same AI you already use — now ready for the work your team actually needs to do.

Talk to sales

Rolling out to a whole team? Talk to sales →