Obiguard
Cybersecurity platform that inspects AI traffic, enforces policy, redacts data, red-teams agents and monitors infrastructure
A cybersecurity platform that inspects AI prompts, responses and tool calls, enforces policy, redacts sensitive data, red-teams agents and monitors logs and vulnerabilities to govern AI use and secure underlying infrastructure. It sells to security and legal teams in B2B organizations that run AI systems, including health tech, financial services, startups, enterprises and public sector entities. It is positioned as one connected solution where Governance AI, a governed chat workspace and a security operations center work together as a single line of defense, built using AI to detect, explain and remediate risk.
Key features
- Inspect every prompt, response and tool-call
- Policy sets, criteria and actions
- Allow-lists for models and tools
- Append-only audit ledger
- 11 guardrail detectors at gateway
- Project Moonshot red-team evaluations
- Nightly ethics and bias scoring
- Coverage across 5 AI frameworks
- Prompt injection and jailbreak defense
- Automatic PII/PHI/PCI redaction
- Shadow AI discovery
- Agentic AI and tool-call governance
- Per-workspace model allow-lists
- SAML/OIDC single sign-on
- Severity-scored alerts across logs and traces
- CVE Radar with upgrade-impact analysis
- Threat intel cross-matched against CISA KEV
- Automatic LLM code review on push