Show HN: TrustVector – Trust evaluations for AI models, agents, & MCP

2 points

4 months ago

We built TrustVector (trustvector.dev for website) because “which model/agent/tool should we trust?” keeps getting answered with vibes, marketing, or outdated benchmarks. And a lot of our enterprise customers kept asking about it.

TrustVector is an open-source evaluation framework + public directory where each system gets a multi-dimensional trust score across: - Security (prompt injection/jailbreak resistance, data leakage) - Privacy & compliance - Trust & transparency (hallucination/bias, documentation quality) - Performance & reliability - Operational excellence

Key idea: every score is evidence-based (sources + confidence), and you can re-weight dimensions CVSS-style depending on your use case.

Current coverage: 100+ evaluations across models, agents, and MCP servers.

GitHub + methodology are linked from the site. I’d love feedback on: 1) whether the dimensions/weighting are sane, 2) what evidence sources we’re missing, 3) What contribution workflow would make this actually community-maintained?

(Also: this project is not affiliated with trustvector.ai.)

1 comment