LoginSign Up
← Back to all sparks
P

Port

INFRA · APIS
Velocity7.5

Internal developer portal for platform engineering, service catalogs, and scorecards.

www.port.io ↗

Port is turning its catalog into the permission layer other apps and GitHub inherit.

developer-platformai-governanceworkflow-orchestrationmcpgithub-integrationidentity
◆Current state
Port now runs governed Workflows with a growing library of native actions across GitHub, GitLab, Azure DevOps, Linear and Jira, and its AI Builder can execute catalog changes behind approvals. September added External Interfaces, which let custom apps sign users in with Port and act under their existing Port permissions, plus a one-way sync of Port properties into GitHub's new external custom properties. The Pages API went GA and Port AI can now route through any Anthropic-compatible gateway.
◆Where it's heading
Port is pushing its data and permissions outward instead of keeping users inside its UI: apps built elsewhere authenticate against it, GitHub Rulesets target repos by Port-managed criticality and ownership, and agents reach it through MCP. Governance is also being decentralized, with scoped building permissions handing blueprint and workflow ownership to individual teams. The monthly cadence of one platform-level expansion plus a batch of workflow actions has held since spring.
◆Prediction
Expect the GitHub property sync pattern to extend to other systems (GitLab, Azure DevOps) and more Workflow actions per release; External Interfaces will likely gain templates or starter apps tied to the Lovable/v0 path the release highlights.

◆Recent moves

  1. 20d ago

    Port Product Updates - What we built in September 2026

    ⚡ SPARK

    September moves Port from a place engineers visit to a layer other tools inherit from: custom apps authenticate with Sign in with Port, and Port properties now drive GitHub Rulesets directly. It extends the governance arc of Workflows GA, with scoped building permissions and a wide set of new native actions.

  2. 1mo ago

    Port Workflows reaches GA with Cursor agent orchestration and MCP permission scoping

    ⚡ SPARK

    Port Workflows reaching GA is the foundational move — it defines the governed orchestration layer that everything else (Cursor agent sync, Slack approvals, MCP permissions) plugs into. Adding Cursor Cloud Agent sync and triggering from workflows means Port can now orchestrate AI agent work alongside human approvals in a single governed flow.

  3. 2mo ago

    Port AI Builder executes real catalog changes; Claude Managed Agents integration ships

    ⚡ SPARK

    Port AI Builder executing real catalog changes — with plan-before-execute and per-tool approval controls — shifts Port from a query interface to an active agent operating on engineering infrastructure. The Claude Managed Agents integration in the same release extends the catalog to include AI agent activity as first-class catalog data.

  4. 3mo ago

    Port Product Updates - What we built in June 2026

    Workflows in open beta with GitLab CI/CD triggers and external MCP server connections expands the orchestration surface significantly — more CI/CD systems covered, more external tools reachable from inside Port AI. The private pages and Slack thread memory round out usability gaps that would slow adoption.

  5. 4mo ago

    Port Product Updates - What we built in May 2026

    The public port-plugins repo opens the extensibility model and is the earliest signal of Port's ecosystem ambition. The multiple workflow triggers and sidebar redesign both reduce friction for teams building complex automation flows, and tracking Claude Code / Copilot usage natively extends Port's AI observability story.

  6. 5mo ago

    Port Product Updates - What we built in April 2026

    Custom Widgets let teams embed their own React/JS components inside Port, and the any-OpenAI-endpoint setting makes Port AI work with self-hosted and Azure-hosted models — two important enterprise flexibility levers. Structured output schemas for AI workflow nodes make agentic automation reliable enough to use in production pipelines.