AIR Blackbox is an open-source reverse proxy that creates tamper-evident audit trails for every LLM API call. Records are HMAC-SHA256 chained — tampering is instantly detectable. Unlike observability tools, AIR focuses on provability: • Cryptographic audit chain for tamper-proof integrity • EU AI Act compliance (20 controls, Articles 9-15) • PII detection and prompt injection scanning • Shadow AI detection with model registry • Works with any OpenAI-compatible API pip install air-blackbox

ElevenAgents by ElevenLabs — Scale conversations without scaling your team

Scale conversations without scaling your team

Hey Product Hunt! 👋 We built AIR Blackbox because we kept seeing the same problem: companies are shipping AI features fast, but nobody can prove what their LLMs actually said. When a regulator, customer, or auditor asks "what did your AI tell that user?", most teams are scrambling through logs that don't exist. AIR Blackbox is a drop-in reverse proxy that sits between your app and any LLM provider. Every call gets recorded as a tamper-evident audit record with HMAC-SHA256 chaining — if anyone modifies a record, the chain breaks. What makes it different from observability tools like Langfuse or Helicone: those answer "how is the system performing?" — AIR answers "what exactly happened, and can we prove it?" It's fully open source (Apache 2.0), takes one line of code to integrate, and includes built-in EU AI Act compliance checks for the August 2026 enforcement deadline. Try it now: pip install air-blackbox && air-blackbox demo We'd love your feedback! 🚀

ElevenAgents by ElevenLabs — Scale conversations without scaling your team

Scale conversations without scaling your team

AIR Blackbox

The flight recorder for AI Agents

The flight recorder for AI Agents