hlido 0.2.0

Hlido — independent, evidence-backed trust scores for AI agents. One package: `pip install hlido` for the core, plus framework extras hlido[nemo|langchain|crewai].

hlido

Independent, evidence-backed trust scores for AI agents — from Hlido.

This is the umbrella package for Hlido's Python tooling. Today it installs and re-exports hlido-trust — drop-in agent-vetting tools for CrewAI and LangChain, plus a zero-dependency client — so you can vet an agent before you delegate to it.

pip install hlido               # core (pulls hlido-trust)
pip install "hlido[nemo]"       # + NVIDIA NeMo Agent Toolkit plugin
pip install "hlido[langchain]"  # + LangChain tools
pip install "hlido[crewai]"     # + CrewAI tools

One package. Everything Hlido offers for Python agents lives under the single hlido name — framework support is an opt-in extra, never a separate package to hunt down.

from hlido import HlidoClient

hlido = HlidoClient()
if not hlido.trust_gate("aider", min_score=70):
    raise RuntimeError("Aider failed the Hlido trust gate — not delegating.")

Full docs, the framework adapters, and examples live with the focused package: hlido-trust.

As Hlido's Python surface grows, the full SDK will live under this same hlido namespace.

License

MIT · Hlido [e-mail] · https://hlido.eu