PromptGuard vs Sanctum Runtime

PromptGuard is a security layer between an application and its LLM provider, inspecting requests before they reach the model. It describes 15 detectors across six layers for risks such as prompt injection, jailbreaks, PII, data exfiltration, toxicity, fraud, secrets, malware and tool injection. The product says it detects and redacts 43 PII types using reversible tokenization. Its SDK can automatically instrument supported LLM calls with one initialization call while leaving existing provider code unchanged. Listed providers include OpenAI, Anthropic, Gemini, Azure, Bedrock, Mistral, Cohere, DeepSeek, Groq, HuggingFace, Ollama and vLLM. Deployment options include managed cloud, hybrid self-hosting and air-gapped installations. In zero-retention mode, prompt and response content is not stored; default security events include a shortened preview and content hash. PromptGuard says customer prompts, completions and documents are not used to train, fine-tune or evaluate models. The Free tier is permanent and includes 20,000 scans a month, one API key, one project and 24-hour log retention. Paid plans include a 14-day money-back guarantee, not a trial.

Sanctum Runtime is middleware that checks actions from AI agents, robots, and automated systems before they are carried out. It evaluates proposed actions against policies and can approve them, ask for verification, or block them. Operators can review, approve, deny, or escalate actions through web and mobile workflows. The product describes behavioral monitoring, audit logs, policy history, and decision trails for runtime visibility and incident review. Its open-core runtime supports local policy checks and offline verification, including local Ollama risk scoring. Documented integrations include OpenAI, Claude/MCP, Gemini-style function calling, DeepSeek, Grok, Bedrock-compatible endpoints, NVIDIA NIM, custom agents, and local runtimes. Developers can use Node and Python SDKs, an MCP wrapper, ROS2-oriented policy packs, and smart-home and embodied-agent templates. The repository is described as MIT-licensed and self-hostable; a hosted console and proxy path are also available. The product identifies itself as embedded runtime middleware, not a standalone consumer app or robot controller. Some enterprise capabilities, including fleet orchestration and organization-wide policy sync, are private.