TrustGuard
Runtime security layer that inspects every request and response between AI agents and their models, MCP servers and tools across gateways, SDKs, browsers and sidecars, applying policies for prompt injection, PII and secrets, toxicity and topics with multi-turn context and an allow, block, transform, report or ask verdict on each interaction.
Find alternatives to TrustGuard
What it does
- Guardrails and safety
Sources
- Pricing
-
TrustGuard has its own rate-card section: "10M processed characters / month" then "$0.30 / 1M processed characters" then custom rates for Enterprise, and "100 inspections / month" at "$0.005 / inspection" for Document & URL inspection; the Free plan is $0 a month, Standard $49 a month as wallet credit.
- Description
-
Official documentation · 21 Oct 2026
"TrustGuard inspects every interaction and stops attacks at the moment of execution — on every surface where your agents run: gateways, SDKs, browsers, and platforms." and "Allow, block or transform in-flight — or alert only, depending on fidelity."
- What it does
-
Official documentation · 21 Oct 2026
"Protocol-typed policies run detection rules on every request and response." (the pricing page https://neuraltrust.ai/pricing lists prompt injection, PII and secret detection, sensitive-data redaction, toxicity and topic moderation rows under TrustGuard).
- Sold within
-
Own rate-card rows and a free self-serve start. FAQ: "Both come with every plan, and your wallet is spent on whatever you actually use." - the schema holds one route, so it is encoded standalone (no parent offering); the shared-wallet billing is stated here.
- Name
-
Official documentation · 21 Oct 2026
"TrustGuard inspects every interaction and stops attacks at the moment of execution" - the page, the header Product menu and the pricing page's TrustGuard column all use the name TrustGuard.
- URL
-
Official documentation · 21 Oct 2026
Product-menu entry "Agent Runtime Security" with the line "Protect agent interactions in real time" and the name TrustGuard, linked to this page; the page is headed by TrustGuard throughout.
- Deployment
-
Official documentation · 21 Oct 2026
Deployment section: SaaS ("Fastest to start. Managed infra, automatic upgrades."), Hybrid cloud ("SaaS control plane, privately hosted data plane. Data never leaves your boundary.") and On-premises / air-gapped ("Full isolation. For defence, government, and financial sector."). The section's lead sentence names TrustGate; it sits on the TrustGuard page and the pricing table lists Private deployment under the shared platform rows.
- Interfaces
-
Official documentation · 21 Oct 2026
Docs: "TrustGuard Python SDK", "TrustGuard Node.js SDK" and "TrustGuard Go SDK"; evaluate API "{TRUSTGUARD_BASE_URL}/v1/evaluate"; the collector API key is created in the console (Agent Security, Collectors, Catalog).
- Modalities
-
Official documentation · 21 Oct 2026
Docs: "Evaluates the interaction and returns" a verdict for each request and response (allow, report, transform, ask or block; a transform forwards a transformed payload); prompts and tool traffic are text, and the pricing page (https://neuraltrust.ai/pricing) lists a Document & URL inspection row (document in).
- Intended for
-
Official documentation · 21 Oct 2026
"Battle-tested inside the most regulated environments." and SDKs for developers (https://docs.neuraltrust.ai/integrations/overview lists the TrustGuard Python, Node.js and Go SDKs).
Also from NeuralTrust
-
TrustGate Infrastructure service
TrustGate routes application and agent calls to models and tools while applying access controls, traffic policies, and guardrails.
-
TrustTest Developer tool
Python red-teaming and evaluation framework that points built-in adversarial probes (prompt injection, data leakage, system-prompt extraction, agent and tool abuse) and functional tests at a model, agent or HTTP endpoint, generates test cases from datasets, RAG sources or prompts, and scores every response with heuristics and LLM-as-judge evaluators, locally or on the NeuralTrust platform.
Open NeuralTrust in the directory
Something wrong here? Send a correction — quote this product id: neuraltrust-trustguard.