NeuralTrust
Barcelona agent security company with a runtime guard and an agent gateway, priced per request from a free tier
Quick Overview
| Company | NeuralTrust, S.L. |
|---|---|
| Category | AI Security Platforms |
| Headquarters | Barcelona, Spain |
| Founded | Not stated |
| EU Presence | Yes - Spain (EU) |
| Data Location | Free plan: US or EU data residency; private deployment on the enterprise plan |
| Open Source | No |
| GDPR Compliant | Yes |
| Self-Hosting | On-premise or hybrid deployment (enterprise) |
| Main Features | Agent gateway (TrustGate), runtime security (TrustGuard), MCP catalogue, model routing, traffic policies, LLM budgets, topic classification, SIEM integration |
| Pricing | Free: 100,000 requests/month. Standard from $49/month (listed in USD) plus usage. Enterprise on request. |
| Best For | Teams that want a gateway and runtime guard they can start on for free |
| Replaces | HiddenLayer AI guardrails and runtime security |
Detailed Review
NeuralTrust splits its offer into two products. TrustGate routes agent traffic to models and tools and applies policy there, and TrustGuard inspects interactions for attacks and data leakage.
The terms of service are entered into with NeuralTrust, S.L., and the privacy policy lists an address in Barcelona. The pricing page is rendered on the server, so the numbers below can be read without a login.
Priced per request
The Free plan allows 100,000 requests a month with five team members, seven days of log retention and a choice of US or EU data residency. Standard is listed at $49 a month with a usage credit, then $0.0003 per gateway request.
TrustGuard security processing is metered per million characters on the same page. Prices are shown in US dollars, which matters for a European budget holder.
Gateway and guard together
TrustGate handles model routing, fallbacks, rate limits, tool permissions and budget caps. TrustGuard adds detection of prompt attacks, with the vendor claiming latency under 100 ms.
Because both sit in the request path, one policy can cover homegrown agents, third-party SaaS agents and endpoint use.
Deployment options
The site describes a split-plane design that can run as SaaS, on-premise or both, so data can stay where policy demands. The enterprise tier adds private deployment, SSO and SAML, and signed MSA and DPA.
Limitations to Consider
All figures on the site, such as 99 percent detection accuracy and attacks blocked, are the vendor's own and not independently tested here. Prices are in US dollars. The free tier retains logs for seven days only. For testing before release rather than runtime enforcement, Mindgard is the closer match.
Who Should Use NeuralTrust
European engineering and security teams that want a cheap way to put a gateway and a runtime guard in front of agents and expand to private deployment later.
Pros and Cons
Pros
- Published free tier and per-request prices
- Gateway and runtime guard in one vendor
- US or EU data residency on the free plan
- On-premise and hybrid deployment
- Spanish company with EU terms
Cons
- Prices listed in US dollars
- Free plan keeps logs for 7 days
- Performance figures are self-reported
- Founding year not stated on pages reviewed
Alternatives to NeuralTrust
Other European ai security platforms worth comparing before you decide:
Mindgard
United Kingdom · London AI red teaming that attacks, fixes and retests
Advai
United Kingdom · UK AI assurance: choose, test and monitor AI systems
Giskard
France · Paris open-core red teaming and testing for AI agents
QuantPi
Germany · Saarbrücken validation engine with statistical confidence
See all European ai security platforms or compare them directly against HiddenLayer AI guardrails and runtime security on our alternatives page.
Frequently Asked Questions
Yes. The pricing page lists a Free plan at $0 with 100,000 requests a month, five team members and seven days of log retention.
The plan lists SaaS hosting with US or EU data residency. Private deployment is part of the Enterprise plan.
TrustGate is the agent gateway for routing, tool access and budgets, while TrustGuard is the runtime security layer that screens interactions for attacks and leaks.