Agent Fabric

Agents everywhere.
Managed under one control plane.

Manage, govern, and scale your multi-vendor AI landscape from a single pane of glass to maximize ROI, increase productivity, and continuously improve agent performance.

Image of Agent Fabric connecting AI tools
Get to know Agent Fabric

Click through an interactive tour to see how Agent Fabric transforms scattered AI tools into a tightly managed and coordinated agent network that enables you to scale with confidence.

Oversee your holistic agent network

Instantly discover agents and tools built anywhere so that they can be centrally managed and governed

Activate governance across AI assets

Manage security, tokens, usage, and cost across agents, tools, and models to reduce risk and increase accountability

Speed up work with intelligent orchestration

Run multi-step agentic workflows with your choice of determinism to ensure predictable and reliable results

Discover

Gain complete visibility into your agent ecosystem. Automatically find, catalog, and understand agents and MCP servers across your enterprise. Build a unified registry for collaboration and reuse, ensuring your teams know what assets exist and how they can be leveraged securely and effectively.

A centralized catalog, built on MuleSoft's Exchange to help teams consolidate every agent and MCP server, whether they're built in-house, embedded in SaaS apps, or sourced externally. By bringing every AI asset into a single place, developers and agents will be able to dynamically discover and reuse them at scale.

Image of a dashboard

Use Agent Scanners to automatically detect and register AI agents and MCP servers running across popular ecosystems such as Amazon Bedrock, Google Vertex AI, Microsoft Copilot Studio, and more directly into Agent Registry. This provides IT teams with real-time awareness of all active agents, ensuring every agent can be orchestrated and governed from day one.

Image of a dashboard

Access a curated list of remote, vendor-supported MCP servers, including Informatica MCPs, directly in Agent Registry. This enables you to control, secure, and monitor how developers and agents interact with these external systems, ensuring they are using and reusing trusted MCP servers rather than needing to start from scratch.

Image of a dashboard
Image of a dashboard
Image of a dashboard
Image of a dashboard

Govern

Protect your enterprise with secure, policy-driven control. Ensure every agent interaction is compliant, authenticated, and aligned with enterprise standards. Use centralized policy management to safely scale AI adoption without compromising security or trust.

Standardizes compliance and token management across agents, MCP servers, and LLMs. AI Gateway enforces cost-control, prompt protection, usage rules, and from a central point, ensuring your data stays safe and your budget stays on track.

Image of a dashboard

Ensure agents perform tasks using specific user permissions rather than broad-access accounts with Trusted Agent Identity. "Just-in-time" authorization keeps every agentic interaction secure and fully auditable.

Illustration of AI Gateway tool.
Image of a dashboard
Illustration of AI Gateway tool.

Orchestrate

Connect and coordinate agents and tools with ease to ensure your digital workforce operates with structure and purpose. Unify every asset within the agentic enterprise, enabling them to collaborate intelligently and deliver consistent, reliable outcomes.

Intelligently engage and guide the best-fit agents and tools to ensure high-fidelity results. With Agentforce Script embedded into Agent Broker, you can define fixed rules and hand-offs while letting LLMs handle the reasoning in between. Now, goal-based, autonomous agents are paired with your trusted, codified workflows.

Image of a dashboard

Easily convert any app or API into a secure MCP server using MuleSoft MCP Support. Use a pre-built connector or enable MCP at scale with MCP Bridge, which transforms multiple APIs into agent-ready tools through simple configuration and without code changes.

Image of a dashboard

Standardize agent communication on A2A protocol with MuleSoft A2A Support. Use a pre-built connector to make any agent A2A-compliant so that agents across ecosystems can collaborate seamlessly.

Image of a dashboard
Image of a dashboard
Image of a dashboard
Image of a dashboard

Observe

See how your agents work — and make them better. Bring transparency to your AI operations by tracking agent reasoning, performance, and collaboration patterns to validate decisions, optimize efficiency, and continuously improve outcomes.

Gain a visual understanding of your agent ecosystem. See a high level map of your entire agent network, while being able to dive into various domains and individual relationships between agents and MCP servers.

Image of a dashboard

Track real-time performance across agents, APIs, and systems. With centralized monitoring and actionable insights, you can quickly detect anomalies, troubleshoot errors, and ensure your digital workforce performs reliably at scale.

Image of a dashboard
Image of a dashboard
Image of a dashboard

Keep up with the latest trends, insights, and conversations.

Frequently asked questions

Agent Fabric is an enterprise agent control plane that enables organizations to discover, govern, orchestrate, and observe any agent, AI asset, and agentic workflow from a single control plane. It includes enhancements to existing products and new product capabilities built on MuleSoft.

Agent Fabric is a solution that includes five main capabilities:

Enterprise actionability: MuleSoft MCP and A2A support make any API and agent usable in your agentic workflows. Additionally, use MuleSoft AI Connectors as well as Intelligent Document Processing, Robotic Process Automation, and Partner Manager to connect any LLM or vector database, as well as unstructured, legacy, and partner/EDI data.

Agent discovery: Agent Registry is a centralized location where AI assets are cataloged, easy to find, and ready to be reused or composed into solutions. It includes a curated list of public MCP servers from the Official MCP Registry and from third parties such as Informatica. Additionally, Agent Scanners automatically find and register agents and MCP servers built on major platforms like Amazon Bedrock and Agentcore, and Google Vertex AI.

Agent orchestration: Agent Broker is an intelligent routing service that orchestrates agents and tools across ecosystems, ensuring they can work across the right resources at the right time. With the adoption of Agentforce Script, Agent Broker supports guided determinism which is the ability to balance fixed rules with LLM flexibility, resulting in more consistent and reliable outcomes.

Agent governance: MuleSoft AI Gateway provides a single control point to manage costs, usage, and security and compliance across all of your agents, MCP servers, and LLMs. Additionally, Trusted Agent Identity helps ensure that agents act with a specific user’s identity and proper authorization to keep every agent interaction secure and auditable.

Agent visualizer: Agent Visualizer provides a visual map of your agent network and associated domains. It shows how agents are interacting and includes metrics like confidence scores, bottleneck detection, and hallucination risks.

Agent Fabric is generally available today with new and enhanced capabilities rolling out each month.

Choose to begin your agentic journey with foundational capabilities like making your data foundation accessible to agentic workflows by using our MCP capabilities. You can also focus on discovering and registering all of your agentic assets so that they can be properly managed and governed. Or build your Agentic Enterprise end-to-end with discovery, governance, and orchestration. Get in touch with your account team to get started.

Agentforce is an Agent Platform where you can build, run, and optimize interoperable agents within Salesforce with full lifecycle coverage, trusted context, and controlled behavior. Agent Fabric is an agent control plane that provides a single pane of glass to register, manage, govern, and observe all of your agents and MCP endpoints. Both are powered by the same engine for defining instructions for how agents handle interactions and how they reason, plan, and act — allowing your agentic workflows to interoperate across different platforms.

+

Esta página está disponible en español

Ver en español