Postman has made Fabric Gateway generally available, aiming to give organisations a central way to govern how AI agents and other systems find and use APIs and tools. The important shift is from connecting agents to services to managing what they can discover and access along the way.
Watch Desk analysis
What happened
Fabric Gateway is a protocol-agnostic control plane for AI agents, large language models and Model Context Protocol servers, according to Postman. The company says it can centrally manage discovery, access permissions and compliance across the lifecycle of those connections. Help Net Security’s report says Postman announced the product’s general availability.
Why it matters
Agents that can use APIs and tools need more than a clever prompt: organisations also need a way to govern what those systems can find and reach. A central gateway could make those permissions easier to manage across different protocols, rather than leaving every connection to its own arrangements. That is the promise; the report does not establish how the controls work in practice or how well they perform.
Our read
This is a practical piece of agent infrastructure, aimed at the unglamorous question that becomes rather glamorous after something accesses the wrong system: who can connect to what? Postman’s announcement makes Fabric Gateway worth watching, but buyers will want specifics on how permissions are set, enforced and audited before treating a control plane as a safety net.
What to watch
- Which protocols and services Fabric Gateway supports in practice.
- How organisations configure and enforce access permissions.
- What compliance and audit features are available.
- Whether Postman publishes implementation detail or customer examples.
Discussion spark: Should organisations route AI agents through a central gateway by default, or does that create a single point of failure and control?
Sources and evidence
- Postman adds security controls for AI agents, APIs, and MCP servers (29 September 2026, 13:40 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.