Show HN: On-prem-deployable voice agent simulation and evals

Category: ai-ml

Tags: voice-ai, llm-evals, agent-testing, on-prem, simulation

Score: 4.7/10 (Innovation: 5, Technical: 4, Documentation: 2, Utility: 5)

Marker is a proposed on-prem platform for simulating and evaluating voice AI agents, aiming to bring versioned, context-preserving testing infrastructure to voice agent development. The README is a manifesto-style blog post rather than technical documentation, so the actual product, architecture, and usage remain largely unknown.

Target audience: AI engineers and backend developers building voice agents

Repository: https://usemarker.ai/blog/hello-world/

View on Hacker News