)
NeuralTrust TrustGate and Merge Agent Handler are both MCP gateways: both connect AI clients to the systems employees and agents need, both expose a catalog of tools behind a single endpoint, and both tie access to an enterprise identity provider. Merge's pitch is governance for employee AI: connect Claude, ChatGPT, or Cursor to thousands of maintained connectors through one IT-managed URL, with access scoped by role and every call logged. For an IT or platform team that wants to say yes to employee AI quickly, that is a clean story, and Merge tells it well on identity, connectors, and data-loss prevention.
Where the two diverge is what the gateway does with the call once it is allowed, and where the gateway runs. TrustGate is the broader platform. It matches Merge on the enterprise identity story and on data masking, and then covers the ground Merge does not: it is an open-source gateway you deploy in your own private or cloud environment, it exports natively to the SIEMs a SOC already runs, and it inspects each tool call for injection and abuse across a whole session rather than only scanning for sensitive data and logging the outcome. Merge governs who reaches which tool and what data leaves. TrustGate does that and secures what actually flows through the call, on infrastructure the enterprise owns.
TL;DR
- NeuralTrust is an open-source gateway you deploy in your own private or cloud environment; Merge Agent Handler is Merge-hosted SaaS. Even Merge's single-tenant option runs on Merge-hosted servers. With TrustGate the gateway runs inside your own perimeter.
- NeuralTrust exports natively to the SIEMs a SOC already runs; Merge keeps an in-product audit trail. TrustGate ships native integrations for Sentinel, Splunk, QRadar, Datadog, and Elastic. Merge documents a dashboard audit trail, not native SIEM streaming.
- NeuralTrust inspects the call for injection and abuse across a session; Merge scans for sensitive data. Merge's DLP is real and useful, but it is data-loss prevention plus logging, not a runtime detection engine that reasons across a whole conversation.
- Both integrate with your identity provider and both do data masking, so those are ties. The decision is everything else that governs the call, and where the gateway runs.
Comparison at a Glance
| Capability | NeuralTrust | Merge |
|---|---|---|
| Built-in runtime security engine | ✅ | ❌ |
| Session memory / multi-turn analysis | ✅ | ❌ |
| Native SIEM export of gateway events | ✅ | ❌ |
| Open source | ✅ | ❌ |
| Identity provider / SSO & SCIM integration | ✅ | ✅ |
| DLP / data masking on tool calls | ✅ | ✅ |
Platform Overview
What is NeuralTrust TrustGate?
)
TrustGate is NeuralTrust's gateway. As an MCP gateway it sits between agents and the MCP servers they call and becomes the single place where tool discovery, identity, policy, security, and observability attach. Teams connect MCP servers from an integrated catalog of more than 200, expose a curated set of tools behind one endpoint, and govern every one of them through the gateway from the moment they are enabled. It is the control layer a platform or architecture team can standardize the whole agent estate on, running in their own infrastructure rather than a vendor's.
Because NeuralTrust is a security company, that control layer comes with something the others source separately: a Security Engine attaches to every route and inspects each tool call inline, enforcing an allow, block, or transform decision before the call reaches its target, and it maintains session memory so it reasons about a whole conversation rather than one call at a time. TrustGate deploys flexibly across private and cloud environments, applies data masking on the traffic it governs, and exports its events natively to the SIEMs a security team already runs.
)
What is Merge Agent Handler?
)
Merge Agent Handler, in its Agent Handler for Employees configuration, is a governance layer for how employees connect AI tools to the systems IT manages. Employees connect Claude, ChatGPT, Cursor, or any MCP-compatible client with a single URL, every action runs under the employee's identity, and IT defines which tools each team or user can access. Merge maintains thousands of connectors and supports importing public MCP servers, and it connects to Okta, Azure AD, or any SCIM-compatible identity provider so access follows the org chart.
Its strengths are identity-driven access and data control. Every tool call is scanned for sensitive data before it reaches a third-party system, with DLP rules that block, redact, or mask flagged content, and every call is logged with the employee's identity, the action, and the outcome. But Merge Agent Handler is a Merge-hosted SaaS product: even its single-tenant option runs on Merge-hosted servers, not infrastructure the customer operates. Its audit trail lives in the Merge dashboard rather than streaming natively to a SIEM, and its runtime protection is DLP scanning rather than a detection engine that inspects calls for prompt injection or reasons across a whole session. It is built to govern employee access and outbound data, not to be the security substrate an enterprise owns end to end.
Built-In Runtime Security
TrustGate carries security inside the gateway. A Security Engine attaches to every route, inspects each tool call and its response inline, and enforces an allow, block, or transform decision before the call reaches its target. It looks for prompt injection carried through a tool result, abuse patterns, and unsafe content in the call itself, not only for sensitive data on the way out. Detection and enforcement are native to the gateway, on every route, by default.
Merge's runtime protection is data-loss prevention. Every tool call is scanned for sensitive data before it reaches a third-party system, and IT can block, redact, or mask flagged content. That is genuinely valuable and it is one of Merge's best features, but it answers a narrower question: is sensitive data leaving? It does not inspect the call for prompt injection arriving through a tool response, and Merge does not document a detection engine that examines calls for attack content or reasons about intent. DLP stops data from going out; it does not stop a poisoned tool result from steering the agent. TrustGate does both, because it inspects the whole call, in both directions, not only the data on the way out.
Session Memory vs Per-Call Logging
TrustGate's foundation is session memory: it treats an interaction as a conversation, not a sequence of unrelated calls, so it can catch a pattern that only appears across turns, such as abuse deliberately spread across several innocuous-looking tool calls. Each call can look acceptable on its own while the sequence is not, and enforcement that only ever sees one call at a time will pass every one of them.
Merge logs each tool call individually: who authorized it, which system it touched, and what happened. That record is thorough for a per-call audit, and DLP applies to each call as it passes, but Merge documents no conversation-level or multi-turn analysis that connects those calls to see a pattern forming. A sequence engineered to keep every single call under the DLP threshold is exactly the case a per-call model does not catch. TrustGate reasons across the whole conversation as it happens; Merge evaluates and logs one call at a time.
Native SIEM Integration
An MCP gateway produces exactly the events a security operations team wants in its SIEM: who called which tool, what was blocked, which policies fired. Whether the gateway delivers those events cleanly into that tooling decides if it fits how an enterprise already runs security.
TrustGate treats this as a first-class capability. It exports and streams its events natively to the SIEMs security teams already run, with support for Microsoft Sentinel, Splunk, IBM QRadar, Datadog, and Elastic, plus OpenTelemetry as the export standard and a generic webhook for anything else. The events flow into the customer's own pipeline in standard formats, so TrustGate becomes one more high-quality source feeding the SOC.
Merge keeps a full audit trail of user actions inside the Merge dashboard. For reviewing activity in-product that is useful, but it is an in-product trail rather than a native SIEM feed, and Merge documents no named connectors for Sentinel, Splunk, or QRadar. Getting Merge's activity into the SOC's tooling means working from that dashboard and building the pipeline yourself, rather than selecting the destination and streaming events in real time. The record lives where Merge runs, not where the security team already works.
Flexible Deployment vs Merge-Hosted SaaS
Deployment is where the two products differ in kind, not degree. TrustGate is an open-source gateway that deploys flexibly in the customer's own private or cloud environment, from managed cloud for speed to hybrid to fully private for defence, government, healthcare, and financial services. The data plane stays inside the customer's perimeter, so sensitive tool traffic never has to leave it, and there is no lock-in at the infrastructure layer.
Merge Agent Handler is Merge-hosted. Customers choose multi-tenant or single-tenant hosting, and normalized data can be streamed into their own storage, but even the single-tenant option runs on Merge-hosted servers rather than a gateway the customer operates. For teams with data-residency or sovereignty requirements at the gateway itself, that is often the first filter, and it is one a Merge-hosted model does not clear. TrustGate lets the enterprise own the deployment; Merge asks the enterprise to route through Merge's.
Final Verdict
Merge Agent Handler is a strong way for IT to govern employee AI: it ties access to the identity provider, maintains thousands of connectors, applies DLP to outbound data, and logs every call. Those are real strengths, and for an organization whose only goal is scoped, audited employee access to SaaS tools, Merge does that job well. But as the gateway an enterprise standardizes on, it leaves gaps: it runs on Merge-hosted infrastructure rather than the customer's, its audit trail stays in Merge's dashboard instead of streaming to a SIEM, and its runtime protection is DLP plus logging rather than a detection engine that inspects calls for injection and reasons across a session.
NeuralTrust TrustGate wins the broader decision because it does not force those trade-offs. It matches Merge on identity and on data masking, and then covers everything Merge leaves out: an open-source gateway the enterprise deploys in its own private or cloud environment, native export to the SIEMs a SOC already runs, and a Security Engine that inspects the whole tool call for attack content across the entire session, not only the data on the way out. Merge governs employee access and outbound data. TrustGate is the one platform a team can run the entire agent estate on, in its own infrastructure, with security built in rather than sourced separately.
Frequently Asked Questions
1. What is the main difference between NeuralTrust and Merge Agent Handler?
NeuralTrust TrustGate is an open-source MCP gateway that deploys in the enterprise's own private or cloud environment, with a Security Engine that inspects the content of each tool call inline, session-level reasoning, and native SIEM export. Merge Agent Handler is a Merge-hosted governance layer for employee AI, strong on identity-driven access, connectors, and DLP. Both connect agents to systems and both integrate with your identity provider, but only TrustGate runs in your infrastructure and inspects the whole call for attack content rather than only scanning outbound data.
2. Does Merge Agent Handler have built-in runtime security like NeuralTrust?
Merge's runtime protection is data-loss prevention: every tool call is scanned for sensitive data before it reaches a third-party system, with rules to block, redact, or mask. That is valuable, but it does not inspect the call for prompt injection arriving through a tool response, and Merge does not document a detection engine that examines calls for attack content or reasons about intent. NeuralTrust's Security Engine inspects the whole call in both directions and enforces inline before it reaches its target.
3. Can Merge analyze abuse spread across a whole conversation?
Merge logs and applies DLP to each tool call individually and documents no conversation-level or multi-turn analysis, so a sequence engineered to keep every single call under the DLP threshold falls outside what a per-call model evaluates. NeuralTrust carries session memory and reasons across the whole conversation, which is what catches abuse designed to stay under the threshold of any single call.
4. Does Merge integrate natively with SIEM tools?
Merge maintains a full audit trail of user actions inside the Merge dashboard, but documents no named native SIEM connectors, so getting its activity into Sentinel, Splunk, or QRadar means building that pipeline yourself. NeuralTrust exports and streams events natively to Microsoft Sentinel, Splunk, IBM QRadar, Datadog, and Elastic, plus OpenTelemetry and a generic webhook, so its data feeds a SOC's existing pipeline directly and in real time.
5. Can Merge Agent Handler be deployed privately or is it open source like NeuralTrust?
Merge Agent Handler is Merge-hosted SaaS. Customers can choose multi-tenant or single-tenant hosting and stream normalized data into their own storage, but even single-tenant runs on Merge-hosted servers rather than a gateway the customer operates, and it is not open source. NeuralTrust TrustGate is an open-source gateway that deploys in your own private or cloud environment, so the data plane stays inside your perimeter.
Related MCP Gateway Comparisons:
- The 6 Best MCP Gateways for Enterprise AI Security in 2026
- NeuralTrust vs Arcade: MCP Gateway Comparison 2026
- NeuralTrust vs Composio: MCP Gateway Comparison 2026
- NeuralTrust vs Runlayer: MCP Gateway Comparison 2026
- NeuralTrust vs TrueFoundry: MCP Gateway Comparison 2026
About the Author
Alessandro Pignati is Lead AI Security Researcher at NeuralTrust, where he leads research on AI and agentic security, advancing techniques to evaluate and secure large language models and autonomous AI systems. He specializes in adversarial machine learning, AI red teaming, LLM security, and AI safety, contributing to the development of secure and trustworthy AI.
NeuralTrust is an AI agent security platform, recognized in the Gartner 2025 Market Guide for AI Gateways and Guardian Agents, and the KuppingerCole 2025 Leadership Compass for Generative AI Defense. Headquartered in Barcelona with ISO 27001 certification.
)
)
)