April brings major new capabilities including structured output schemas for AI workflow nodes, support for connecting any OpenAI-compatible LLM endpoint to Port AI, and automatic Engineering Insights provisioning out of the box. Here's everything that's new in Port this month.
Check out the interactive version of these release notes, click here.
Big
Custom Widgets: Build and run your own UI inside Port
Create Custom Widgets to run your own frontend components directly inside Port dashboards and entity pages. Start by develop your plugin in your preferred stack, build and upload it with the Port Plugins CLI, then configure a Custom Widget to use it. Once added, your plugins are available in the Plugins Manager and can be reused across dashboards and entities - fully integrated with Port's data, permissions, and rich context.
For example, you can build a calendar Custom Widget that visualizes incidents over time, and reuse the same plugin to display PRs by simply changing its configuration.
Connect any OpenAI-compatible endpoint as a Port AI provider
Route Port AI through internal gateways, self-hosted models, or any endpoint implementing the OpenAI's API.
Bring your own infrastructure to Port AI and connect any OpenAI-compatible endpoint for full control over privacy, cost, and model selection. Works with LiteLLM, self-hosted models, and any endpoint that implements the OpenAI chat-completions format.
Medium
Access recently viewed items directly from Spotlight Search
Spotlight Search (cmd+k) now surfaces your 20 most recently viewed items for instant navigation back to recent work.
Jump back to any recently visited page, blueprints, entities, actions, etc. without searching from scratch, so you can return to recent work in a single click.
Filter and visualize by Calculation Properties
Calculation properties are now persistent and indexed, delivering ~10x faster entity tables and enabling filtering, querying, and dashboard widgets on top of calculation properties. Combined with the already-supported automation triggers, calculation properties now offer a full suite of capabilities across the platform.
Define Structured Output Schemas for AI Invocations
Add a JSON Schema to your AI workflow nodes or API calls and get validated, structured output backβready to use in downstream nodes without parsing.
Make AI-powered automations reliable by defining exactly what structure the AI should return. Add an
outputSchema
field to your AI or AI Agent workflow nodes and Port AI will validate the response against your schema, and return a structured JSON object you can pass directly into upsert entity or webhook nodes.π Learn more in our docs, π Learn more in our API docs
Collect feedback on AI invocations in Port
AI quality is hard to improve when you can't measure it. Users can now rate any PortAI conversation in Port with a thumbs up or down and an optional comment.
Platform engineers get a real signal on what's working, can spot patterns in failures, and have the data to tune prompts and skills over time. Feedback is also available via API for teams who want to pipe it into their own workflows.
Get Engineering Insights Out of the Box
Port now provisions a full Engineering Insights setup automatically during onboarding.
Start measuring engineering performance from day one, no manual configuration required. Connect your tools and Port automatically provisions four dashboards: Delivery Performance, Pipeline Reliability, DORA Metrics, and Scorecards & Standards; plus the blueprints and scorecards to tie services, deployments, incidents, and on-call context together.
Existing customers can adopt the same setup using the engineering insights guides.
Filter Integration Event Logs by Level, Kind, and ETL Phase
New filters in the integration event log let you quickly pinpoint specific entries without scrolling through noise.
Spend less time hunting through logs and more time resolving sync issues. In the data source modal's Event Log tab, select a sync and use the filter popover to narrow results by log level, kind, ETL phase, or message content.
Upload images to Markdown widgets
You can now directly upload images in markdown widgets and markdown properties. Copy-paste, drag-and-drop, or use the toolbar to embed images with no external hosting needed.
Perfect for enriching service READMEs, runbooks, and team documentation with screenshots, architecture diagrams, or visual guides.
Use custom user properties and relation references in markdown widgets
Markdown widgets now support custom user properties and can reference markdown properties from related entities.
Build richer, more dynamic content by accessing custom user fields and pulling markdown from related entities, all using the
{{ .user.properties.field }}
and {{ .entity.properties.field }}
template syntax.To create personalized dashboards, context-aware documentation, and multi-layered content in a single widget.
Ask AI in context from any dashboard or widget
The AI side chat now opens preloaded with context from the dashboard or widget you're currently viewing.
Get AI answers tailored to exactly what you're looking at, without losing your place. Click the Ask AI button on any supported dashboard or widget to open the side chat with that page's data already in scope. Port surfaces suggested questions based on your current view, and you can remove context at any time via the context chip.
Small
Filter dashboards by Entity Creation Date and Last Update
Entity Creation Date
and Last Update
are now available as dashboard-level basic property filters, letting you scope all
widgets to a time range at once
.Scope your entire dashboard to a specific time range with a single filter - no need to configure each blueprint in the dashboard separately.
Grant AI Access to All Members with One Click
Admins can now enable AI access for their entire organization with a single toggle instead of manually configuring blueprint permissions.
Give your entire team access to Port AI in seconds, without manually editing blueprint permissions. Go to Organization Settings β AI tab and toggle on AI access for everyone, or respond to the inline prompt in the AI sidechat when members lack access.
Set the display format for date time properties
Choose how dates appear - relative, 12-hour, or 24-hour, full time stamp, using a new format dropdown in date/time property settings, to pick your preferred display.
Display null values as zero in line chart widgets
A new toggle lets you show null values as zero on line charts while preserving the actual null value in the tooltip.
Keep your line charts continuous and readable even when data has gaps. Enable the 'Show null values as zero' toggle in the Line chart widget form to fill gaps with zero, hover over any point to see the original value in the tooltip (whether it's null or not).
Switch regions directly from the login and signup pages
Your current region is now visible on the login and signup pages, with a one-click option to switch before you sign in or create an account.
Click the region indicator on the login or signup page to switch to the correct region before you sign in or create a new account.
Apply account rate limiting
Port now enforces per-customer rate limits on the entity upsert endpoint, ensuring one customer's traffic can't degrade the platform for everyone else.
Requests to
POST /v1/blueprints/:blueprint_identifier/entities
are limited to 10,000 per 5 minutes (burst: 100 req/s). Customers who exceed the limit receive a 429
response β implementing retries with exponential backoff is the recommended mitigation. No customer currently crosses this threshold, and proactive outreach will be made to any who approach it ahead of the May 21 enforcement date.Integrations Improvements
GitHub Copilot integration - support Enterprise endpoints
You can now integrate enterprise-level usage metrics around GitHub Copilot.
New AWS self hosted integration
Self-host the AWS v3 Ocean integration across multiple AWS accounts to sync AWS Organization resources into Port.
The new installation docs include step-by-step CloudFormation instructions (with parameters and verification steps) for ECS (Fargate), EC2, and IRSA (EKS) deployments, plus a guided, interactive troubleshooting flow to diagnose and fix common issues like connectivity, auth, OOM crashes, and StackSet failures.
πͺΒ Built from your feedback
Github Integration - Support for Affiliation Selector in Collaborator Kind
Added support for ingesting and relating Github-affiliated entities (repos, orgs, users)Β within the Github Ocean integration.
πͺΒ Built from your feedback
Snyk Integration - Added ability to filter by project
Added a project filter to the Snyk integration selector, allowing users to restrict ingestion to one or more specific Snyk projects.
πͺΒ Built from your feedback
GitHub Integration - Added support for identifying first commit and timestamp associated with Pull Request
Added support for identifying the earliest commit & timestamp that belongs to a pull request across GitHub, GitLab, and Azure DevOps.
πͺΒ Built from your feedback
OIDC / Workload Identity authentication for Azure Resource Graph
The Azure Resource Graph integration now supports OIDC (workload identity) authentication in addition to client secret credentials.
This allows customers running self-hosted deployments on AKS to authenticate without storing long-lived secrets by leveraging on Azure's native workload identity federation instead.
Under the hood, the integration now uses Azure's DefaultAzureCredential, which automatically detects the available authentication method in your environment (workload identity, managed identity, environment variables, CLI, etc.)
πͺΒ Built from your feedback
GitHub Integration - Access additional data from Github users API
Expand the Github user kind to include additional fields available from the GitHub Users API (e.g. company, bio, location, social handles).
πͺΒ Built from your feedback