Proxiq
Self-hosted LLM gateway with DLP, SSO & full audit trail
Proxiq provides a self-hosted, open-source gateway for large language models, ensuring security and compliance for organizational data. Key capabilities include:
* Data Loss Prevention (DLP) for sensitive information
* Single Sign-On (SSO) integration
* Configurable security policy enforcement
* Comprehensive audit trail of all interactions
* Zero code changes required for developers
This platform allows security teams to monitor and control all communications with external language models like Anthropic, OpenAI, Azure, Groq, and Ollama. It blocks or redacts PII, detects prompt injection attacks, and locks system prompts to maintain compliance. Organizations can define per-user or per-team security policies, ensuring data governance across departments such as banking, healthcare, or marketing, all while maintaining a full audit trail.
For developers, Proxiq offers seamless integration by simply redirecting existing SDKs to a local port (e.g., `localhost:3099`). This eliminates the need for code modifications, ensuring a smooth transition. Additionally, Proxiq delivers significant cost reductions through semantic and exact caching, along with intelligent routing that automatically selects the most cost-effective model for each request.
Proxiq is ideal for enterprise security teams seeking robust control over interactions with external language models and developers who require secure, compliant access without altering their existing development workflows. It supports various identity providers, including Microsoft Azure AD, Google Workspace, GitHub, Okta, and any OIDC/SAML IdP, and ensures no data leaves your infrastructure.