Port Product Updates - What we built in February

In February, we made Port AI more connected and more controllable. A new Azure Anthropic Provider lets you . AI Agent prompts now support up to 5,000 characters, and Port AI is without switching screens.

On the platform side, Dynamic RBAC permissions are now manageable through a dedicated UI, so you can . A new Permissions Simulator lets you . Secrets now have full audit logs so you can . And dashboard filters are now available to all members, so teams can .

These are only the first few of all the great thing we released last month! Check out the interactive version of these release notes, click here.

Manage blueprint permissions through a new dedicated Dynamic RBAC UI — no more editing JSON files. Configure who can read, create, update, or delete entities using role-based, user-specific, team-based, and dynamic ownership rules directly from each blueprint.

We also introduced a brand-new Permissions Simulator. You can now simulate whether a specific user has access to a given blueprint or entity, see if the action is allowed or denied, and understand which policy determined the result — before your changes impact real users.

The simulation function is also available in our public API, allowing you to programmatically check permissions on your data model.

💪 Built from your feedback

🔗 Learn more in our docs

Port now logs create, update, rotate, and delete events for secrets so you can trace unexpected changes and maintain accountability.

Get full visibility into who's changing your secrets and when—critical for security audits and incident response. Navigate to Builder → Audit Log → Secrets to view a complete history of create, update, rotate, and delete events, including who made each change and exactly when.

💪 Built from your feedback

Port's MCP Server now uses consolidated tools that cut context window usage in half, enabling faster AI responses and lower token costs.

Cut your AI's context window usage in half with Port's streamlined MCP tools, leaving more room for meaningful work. The consolidation is automatic—create/update merged into upsert operations, multiple get methods unified into smart list operations. If your agent prompts reference specific tool names directly, update them to the new consolidated names.

: Breaking change — agent prompts referencing old tool names (create_, update_*) must be updated.

🔗 Learn more in our docs

:

🔗 AI API interaction doc

Access Port AI instantly from spotlight search without navigating to a separate interface.

Get AI assistance without breaking your flow—Port AI is now accessible directly from spotlight search. Open spotlight search and type your question. Port AI responds instantly without switching screens or opening new tabs.

AI Agent prompts now support up to 5,000 characters, and AI Invocations expose a new source property for filtering by origin.

Build more sophisticated AI Agents with twice the prompt capacity and greater control over how invocations are tracked. Write AI Agent instructions up to 5,000 characters—double the previous limit. Use the new property on the AI Invocation blueprint to filter invocations by origin across dashboards, views, and automations.

Use Anthropic Claude models through your Azure AI Services account with the new Azure Anthropic provider.

Bring your Azure-hosted Claude models to Port AI while keeping all traffic within your Azure environment. Configure the Azure Anthropic provider via API with your Azure API key, resource name, and model deployments mapping. Supports Claude Sonnet 4.5 and Haiku 4.5 deployed on your own Azure infrastructure.

: API-only configuration, no UI yet.

🔗 Learn more in our docs

Integrations now run more reliably at scale. We increased resources for high-usage hosted-by-Port integrations (like Jira, ADO, Github, Pager duty and more) and added production-based CPU and memory recommendations to help you size self-hosted integrations and avoid OOM issues.

Check the updated integration docs to review the recommended resource sizes.

🔗 Jira docs, 🔗 GitLab docs, 🔗 Azure DevOps docs, 🔗 PagerDuty docs, 🔗 GitHub Ocean docs, 🔗 AWS docs, 🔗 Terraform docs

Dashboard filter management is now available to all users, not just Admins.

Team members can now customize dashboard views on their own. Members can add, edit, and remove filters on dashboard pages, filtering by team, environment, or any other criteria. These changes apply only to their personal view, so they can explore data without affecting how others see the dashboard or waiting for an admin to update it. At any time, members can revert back to the default filters configured by the admin(s).

💪 Built from your feedback

🔗 Learn more in our docs

Port AI now guides you through setting up dynamic permission policies for self-service actions.

Let AI help you configure complex permission rules for self-service actions—no policy syntax knowledge required. Ask Port AI questions like "Help me set up dynamic permissions for my action" or "How do I prevent self-approval?" to get guided JSON configuration for execute and approve policies.

Action widgets opened from entity pages now automatically pre-select that entity, removing a redundant step.

Skip manual entity selection and run actions faster when launching from a specific entity page. Add a day-2 or delete action widget to any entity page—when opened from that page, the entity is already selected and ready to go.

💪 Built from your feedback

The top navigation options menu has been reorganized for quicker access, with 'View as' now surfaced directly.

Find what you need in the top navigation faster with a reorganized options menu. The '...' overflow menu is removed—'View as' and other common actions now appear directly in the main options menu, reordered for more intuitive discovery.

The Azure DevOps Ocean integration now supports webhook events for Pipelines, Pull Requests, Repositories, and Work Items.

Track more than just code changes with expanded webhook event support in the Azure DevOps Ocean integration. The integration now responds to Pipeline, Pull Request, Repository, and Work Item events, giving you real-time visibility into builds, code reviews, and project management activities.

💪 Built from your feedback

🔗 Learn more in our docs

You can now sync Checkmarx applications into your Port catalog to create a clear application-level context for your security data. This helps connect related projects under a single application boundary, improving traceability and making governance, reporting, and risk management easier.

Add the Checkmarx integration from to start syncing.

💪 Built from your feedback

🔗 Learn more in our docs

You can now ingest custom fields from PagerDuty services into your Port catalog. This lets you use existing service metadata, like team, environment, or internal IDs, to model relationships and enrich your service context without duplicating data.

Enable in your PagerDuty integration and map fields to blueprint properties using .

💪 Built from your feedback

🔗 Learn more in our docs

New implementation guides are now available for the autonomous ticket resolution workflow, covering the full ATR solution.

Get started with autonomous ticket resolution using a complete set of implementation guides, from overview to hands-on patterns. Start with the new solution overview, then follow guides for work item blueprint patterns, AI-powered work item preparation, and calculating blast radius with AI.

🔗 Learn more in our docs

:

An update regarding our data processing activities - Brizz will become a subprocessor for Port. Brizz is a product analytics solution, focused on agentic systems (such as LLM agents, AI copilots, AI support bots) which will allow us at Port to continue creating quality products. We take data privacy and security seriously, and this change reflects our commitment to maintaining the highest standards in protecting your data.