March brings two major additions to Port:
MCP Connectors
plug external MCP servers directly into Port AI
, turning it into a unified gateway for your entire tool stack, and Visual data source mapping
(beta) replaces hand-written YAML for configuration with guided, form-based interface. On the platform side, multi-line charts, baseline reference lines, and saved drill-down views make dashboards more powerful and consistent—and revamped engineering intelligence guides help you
put those metrics to work
.Big
Connect External MCP Servers Directly to Port AI
Connect external MCP servers directly to Port AI, giving your team a single interface to tools from Atlassian, Notion, GitHub, and more.
Turn Port AI into a
unified MCP gateway
for your entire tech stack by connecting external services directly to the side chat. Org admins add an MCP server and select which tools to expose—users can then access them through Port AI chat in the portal and Port's own MCP server, making Port the single entry point for AI-powered interactions across your organization.
Visual data source mapping (Beta)
Configure data extraction and ingestion visually using dropdowns and forms instead of hand-written YAML.
Eliminate the back-and-forth of manual YAML configuration by mapping your data sources directly in a guided, form-based interface. Select your kinds, selectors, and blueprint properties from dropdown menus, then validate everything with the built-in Test button per each resource.
Medium
Multi-line chart
Combine multiple metrics—across blueprints and types—into a single line chart with a shared time axis. Compare signals side-by-side (e.g., service count vs. deployment frequency) to better understand platform health.
→
All newly created charts use the multi-line model
. Existing charts remain fully functional and backward compatible.Migration
- Legacy line chart will be deprecated in mid-July 2026.
- IaC users: please follow the migration guide.
Duplicate Any Dashboard Page
You can now duplicate an entire dashboard—including all widgets, layout, and permissions—instantly. Stop rebuilding dashboards from scratch by duplicating any existing one. Open the dashboard menu, select duplicate, and get an exact copy with all widgets, layout configuration, and permissions already in place.
Small
Enrich Git entities with a unified JQ-based mapping syntax
File and search enrichment in git integrations now follow the same JQ-based pattern used everywhere else in Port - no more special
file://
or search://
syntax. Legacy syntaxes remain supported during migration.Custom icons are now GA
Custom icons are now generally available - every organization can now
upload
, use
, and delete
custom icons across the portal.
Configure AI Agent Allowed Actions with a New Dropdown
Define exactly which actions your AI agents can run — and keep full control over their scope.
When setting up an AI Agent, use the new
Allowed Actions
dropdown to select permitted actions directly from a searchable list. Instead of writing regex patterns to configure agent tools, you pick the actions you want to allow and Port handles the rest.Need more precision? Advanced mode lets you write regex patterns to match tools by name — useful for dynamic rules like allowing all GitHub actions, or restricting agents to read-only operations.
Track Next Month's Seat Count After User Deletions
A new Usage Dashboard widget shows your predicted seat count for next month, updated within 24 hours of deleting or disabling users.
---
Configure Units in Number Chart's Display Single Property Mode
The Display single property chart type now supports pre-defined and custom units, matching the other number chart modes. Keep your dashboards consistent by configuring measurement units across all number chart types.
Edit Markdown Entity Properties Directly from the Widget
You can now edit Markdown entity property values inline from the widget editor, without leaving the page.
Add Baseline Reference Lines to Line Chart Widgets
Track targets and thresholds in your line charts with custom named reference lines.
See at a glance how your metrics measure up against targets, thresholds, or benchmarks. Open the line widget editor, add one or more baselines, and define a name, value, and color for each—reference lines appear directly on your chart.
Trigger Catalog Discovery Through MCP
Discover and map catalog entities by simply asking the Port AI Agent or any MCP-compatible client.
Ask the agent to discover entities for a blueprint and it will suggest matches for relations or updates to existing ones. Results open automatically in the Discoveries tab so you can review and approve them right away.
Simulate Blueprint Permissions Through MCP
Check who can access what in natural language, directly through the Port AI Agent or any MCP-compatible client.
The
simulate_blueprint_permissions
tool returns a structured breakdown of roles, team membership, ownership, and policies—for any user, operation, and entity. No more manually checking access one by one in the UI.
Control which blueprints appear in Spotlight search
Admins can now control exactly which blueprints appear in Spotlight search results.
Reduce search noise by setting an organization-level default to include or exclude blueprints, then override it for specific blueprints as needed. Manage these settings from the
Search
tab in Organization Settings, or programmatically via the API and Terraform.This gives teams precise control over which entities are discoverable—helping users find the most relevant results, faster.
Save Drill-Down Table Views for Charts
Admins can now save customized drill-down table views for pie, bar, and number charts.
Adjust column order, show or hide properties, and apply filters—then save these settings per widget. Saved views ensure your team can return to the same structured data view anytime, making analysis more consistent and aligned with your workflows.
Guides
We revamped our guides around engineering intelligence and performance metrics.