NeuralTrust

NeuralTrust sells three connected products. TrustGate is an AI/MCP/A2A gateway — a single Go binary that routes LLM and agent traffic across 9+ providers (OpenAI, Anthropic, Azure, Bedrock, Vertex, Groq, Mistral, DeepSeek), with per-consumer auth, rate limiting, semantic caching, and an MCP aggregation plane so agents reach many tools through one point (source: TrustGate README, github.com/NeuralTrust/TrustGate). TrustGuard sits inline and inspects prompts/responses for jailbreaks, PII, and tool abuse, claiming <100ms added latency at p95 and 20k+ requests/second/node under the vendor's own load tests — these are vendor-reported benchmarks, not independently verified. TrustLens is posture management (agent discovery/inventory) and TrustTest is a code-first red-teaming/eval framework you can run in CI. Deployment options span SaaS, hybrid (control plane SaaS, data plane on your infra), and on-prem/air-gapped.

TrustGate (github.com/NeuralTrust/TrustGate) is genuinely Apache-2.0 licensed, not source-available-with-a-catch. As of this check it has 6 GitHub stars, 3 forks, and 18 open issues, with commits and releases (latest v0.44.1) landing as recently as 27 Aug 2026 — actively maintained but with a very small external community; you're effectively adopting NeuralTrust's roadmap, not a broad open-source ecosystem. The org also publishes plugin repos for Claude Code, Cursor, and Codex, and a Helm chart for self-hosting the full platform.

NeuralTrust's research team is a genuine differentiator: it published the "Echo Chamber" multi-turn jailbreak technique and the "Semantic Chaining" multimodal attack vector, both incorporated into the OWASP AI Security Project taxonomy (tech.eu, 17 Jun 2026, citing the funding announcement). Independent outlets (cybersecuritynews.com, infosecurity-magazine.com) have reported on NeuralTrust's own jailbreak demonstrations against GPT-5 and Grok-4 shortly after each model's release — these are NeuralTrust's disclosed findings being covered as news, not third-party validation of NeuralTrust's product, but they show the team is doing real adversarial research rather than just marketing around it.

NeuralTrust raised a $20M seed round announced 17 June 2026, led by Alstin Capital, with VentureFriends, Seaya, Kibo Ventures, Banco Sabadell, EA Ventures Plug and Play Fund, and Finaves (IESE Business School's VC fund) participating, plus support from the European Innovation Council and Spain's AEI (tech.eu, 17 Jun 2026). Founders are Joan Vendrell (CEO), Victor Garcia (CTO), and Alejandro Domingo (COO); the company is Spain-based. The public team page lists roughly 24 people across engineering, GTM, design, and finance — a real team, but still an early-stage company: a buyer betting on multi-year platform continuity should weigh that against the credibility of a $20M round nine months old at time of writing.

There is no pricing page — neuraltrust.ai/pricing returns a 404 — and every product page's only calls to action are "Get a demo" and "Contact." The only pricing-adjacent detail found anywhere is TrustGate's own /pricing docs, which describe how the gateway meters and reports LLM API cost pass-through (per-token USD rates, discounts, cache pricing) — that's a cost-accounting feature for tracking what you spend with model providers, not what NeuralTrust itself charges. A buyer cannot get even a floor number without booking a sales call.

NeuralTrust's own docs (docs.neuraltrust.ai/neuraltrust/security/overview) state SOC 2 Type II, ISO 27001, and GDPR compliance, and list HIPAA/PCI DSS/FedRAMP as frameworks the platform's controls "help meet" (weaker language than "certified for"). A trust-center portal exists at trust.neuraltrust.ai but requires account sign-in to view — so the underlying SOC 2 report or ISO certificate is not independently checkable without becoming a customer or prospect. Trust these as vendor-attested until you can pull the actual report during procurement.

No G2 or TrustRadius review pages with visible content were found during this check — G2's NeuralTrust page returned a bot-challenge with no reviews rendered. NeuralTrust has been named a Sample Vendor in Gartner's Hype Cycle for AI Governance Technologies 2026 and Hype Cycle for Infrastructure Security 2026 (per NeuralTrust's own news posts; Gartner's inclusion criteria for "Sample Vendor" is not the same as a Magic Quadrant ranking or endorsement). The customer logos shown on the site (Sabadell, Iberia, Abanca, ISDIN, Air Europa, Capgemini, NTT Data) are vendor-displayed and unverified independently in this check.