In December, we made it even easier to get data into Port and keep it flowing smoothly. Our brand new AI-powered integration assistant now troubleshoots failures and suggests mapping fixes. Also, custom integrations can now run fully hosted by Port! No more overhead!
Once your data is in Port, it can transform into something even more interactive. Drill down from pie chart slices directly to the entities behind them. Or display it in a different format altogether like a bar chart! Markdown widgets can now display user and entity properties or even display an entire markdown property. iframes also adapt to dynamic content too!
And we almost forgot, we shipped Scorecards 2.0 - scorecards are now blueprints.
Of course, we shipped tons of other features like dashboard filter improvements, a new, useful user status, a portal feedback experience, and much more!
To see an interactive version of these release notes, click here.
Big
AI-powered integration assistance in Port
Port’s integration UI now includes AI-powered assistance to help you fix failed syncs and improve mapping configurations faster. Use "✨Troubleshoot" on the Sync Status tab or "✨Ask AI" in the Mapping tab to get instant suggestions based on your current setup.
Available automatically for AI-enabled organizations, just open an integration and try it out.
💪 Built from your feedback](https://roadmap.port.io/ideas/p/ai-assistant-for-mapping-entities)
AI-powered catalog auto-discovery
Port now uses uses AI to analyze your existing blueprints and their relations to propose new or enriched catalog entities. Review, edit, and approve suggestions; configure related blueprints and add context to improve results. Requires related blueprints and create/update permissions; processes up to 500 recent entities per source; available with the AI_ELIGIBLE feature flag.
Scorecards 2.0
Scorecards are now first class catalog entities in Port, with native blueprints, a dedicated UI, and built in dashboards. Teams can define standards, manage rules, track results, and remediate issues in one place, turning scorecards from static measurements into actionable standards. We are gradually rolling this out to users.
Medium
Port-hosted Ocean custom integrations (Open Beta)
You can now create and run Ocean Custom integrations fully hosted by Port. No Helm charts, Docker containers, or scripts required. This makes it super easy to connect tools not included in the integrations library using a simple, UI-driven setup.
Head to the Data Sources page and add a new Ocean Custom integration to get started.
Pie chart drill-down to underlying data
You can now drill down into Pie chart segments to explore the entities behind each slice.
Tooltips now also show the percentage for each slice, so you can understand the split at a glance.
Filter dashboards by any property
You can now filter dashboards by any property and apply those filters across multiple widgets at once. This lets you explore data consistently without having to configure filters on each widget separately. Also supported in IaC.
New bar chart widget
You can now add bar chart widgets to your dashboards. Use it to compare values across categories, similar to the pie chart widget.
Markdown widget displays dynamic user and entity data
Add contextual variables to Markdown widgets, so content can automatically adapt to the current user and, on entity pages, the current entity.
Use template values like
{{ .user.firstName }}
, {{ .entity.title }}
to personalize what is shown.
Markdown widget can now bind to entity markdown properties
You can now use the Markdown widget to display dynamic content from your entities.
Instead of static text, the widget can bind to a Blueprint, a specific entity, and a Markdown type property. This keeps Markdown content on dashboards and entity pages automatically in sync with your entity data.
- On dashboard pages, you can connect the widget to any Blueprint → entity → MD property.
- On specific entity pages, it always reflects the current entity and you set the MD property.
Existing Markdown widgets remain unchanged unless you switch their source to a Markdown property.
See the form and an example of the rendered widget:
Add users to Port without sending an invitation email
Admins can add now users to Port organizations in 'Staged' status via API or provisioning. That means that the user gets added as an entity but they won’t receive an invitation emails. When they’re ready, users can log in (including via SSO) and they’ll automatically be moved to 'Active' on first login. No configuration changes or breaking changes; available now to all customers.
Portal feedback experience
Easily capture user feedback with the new Portal Feedback experience. Submit issues, request new solutions, upvote ideas, and track trends across your organization — all from structured, purpose-built dashboards. It’s a seamless way to turn user input into meaningful improvements.
Go to
Catalog → +New → New Experience → Portal Feedback
to set it up.Small
Display formatting for number widget
You can now control how numbers are displayed in the Number widget. Choose rounding or customize a specific number of decimal places, like .0, .00, .000, or more, to match your reporting needs. This helps keep dashboards consistent and easier to read at a glance.
Port AI discovery prompt via MCP
Adds a new prompt to help users identify Port AI/MCP capabilities, improving discoverability and understanding. Available through the MCP server and within Port AI.
AI sidechat stays active next to modals
Sidechat now remains open and interactive while modal dialogs are active, enabling continuous Q&A, query generation, and data validation within modal-based flows.
Use dynamic context in iFrame URLs
You can now interpolate user and entity context directly into iframe widget URLs.
Use template variables like
{{ .user.email }}
or {{ .entity.identifier }}
to dynamically construct iframe links that adapt to the current dashboard or entity page. This makes it easy to embed external tools, such as filtered Grafana dashboards, that automatically reflect the relevant user or entity.
Icon picker visual upgrade
We improved icon clarity, grouping, and layout for faster discovery and a cleaner browsing experience.
Unified MCP action execution
We consolidated all action execution into a single
run_action
tool. This reduces the overall tool count and improves context-window efficiency when working with Port AI.Allowed actions header
We also introduced the x-allowed-actions-to-run header, which allows you to allowlist specific actions for execution. Passing an empty string blocks all actions. This change is fully backward compatible and available now across Port AI, the UI, and MCP clients.
MCP Oauth token refreshes automatically
You no longer need to re-authenticate every session. Your token now lives and refreshes for seamless integration with Port.
Read-only mode for Port MCP
Pass a header to enable read-only mode for the MCP to ensure safer operations focused on intelligence and self-service experience.
ServiceNow Integration OAuth 2.0 Support
You can now connect ServiceNow to Port using OAuth 2.0. This provides a more secure, enterprise-ready alternative to Basic authentication while syncing incidents, users, groups, and catalogs.
Use GitHub GraphQL API for richer pull request data
You can now switch to GitHub’s GraphQL API in your integration config for more detailed and flexible pull request data, including new fields like commits, additions, deletions, review status, and more!
Set
api: "graphql"
in your pull request selector to get started. New integrations
You can now ingest monitoring, incident management, AI usage, and automation data into Port using new Custom Ocean integrations for Prometheus, Azure Monitor, Rootly, OpenAI, and n8n. These integrations help your Platform, SRE, and FinOps teams visualize critical data-from alerts to workflows-right inside your software catalog.
All integrations are available today for self-hosted Port users. Follow the individual setup guides to get started.
Guides
AI workflows
Orchestrate incident response with AI
Automate your entire incident response lifecycle by combining AI agents with PagerDuty, Jira, and Slack to reduce MTTR.
Auto-generate GitHub workflow documentation
Automatically generate and maintain human-readable documentation for GitHub Actions workflows whenever they're created or updated.
Building Port
These guides show how to use Port's MCP server to build and configure Port through natural language conversations directly from your IDE.
Build your software catalog with MCP
Build actions and automations with AI
Define scorecards with AI
n8n Workflows
Automatically resolve tickets with n8n and Port context
RBAC for AI agents with n8n and Port
Implement role-based access control for multi-agent systems, enabling dynamic tool access based on user roles and permissions.
Thank you for reading!
The Port team