director-ai

Real-time LLM hallucination guardrail — NLI + RAG fact-checking

Brought to you by: anulum

Downloads: 0 This Week

Last Update: 2026-03-29

Add to Preferred Sources

Director-AI is a real-time LLM hallucination guardrail that catches factual errors mid-stream before users see them.

Core capabilities:
- Token-level streaming halt: stops generation the moment hallucination is detected
- NLI scoring via DeBERTa/FactCG: 75.8% balanced accuracy on 29,320 LLM-AggreFact samples
- RAG knowledge grounding: vector store integration with Chroma, FAISS, Pinecone, Weaviate, Qdrant, Elasticsearch
- 14.6ms GPU latency (GTX 1060, ONNX), 0.5ms/pair on L40S FP16
- Rust-accelerated BM25 (10.2x speedup) and signal processing

Integrations: LangChain, LlamaIndex, LangGraph, CrewAI, Haystack, DSPy, Semantic Kernel, SDK Guard (OpenAI/Anthropic/Bedrock/Gemini/Cohere)

Domain presets: medical, finance, legal, customer support, summarisation, creative

Security: Sigstore-signed, SLSA provenance, OpenSSF Best Practices (100%), AGPL-3.0 with commercial licensing

pip install director-ai
https://github.com/anulum/director-ai

Features

Real-time hallucination detection, Token-level streaming halt, NLI scoring (DeBERTa/FactCG), RAG knowledge grounding, 8 framework integrations, Domain presets, ONNX/Rust acceleration, Sigstore-signed releases

Project Activity

See All Activity >

Follow director-ai

director-ai Web Site

Other Useful Business Software

Go from Code to Production URL in Seconds Icon

Go from Code to Production URL in Seconds

Cloud Run deploys apps in any language instantly. Scales to zero. Pay only when code runs.

Skip the Kubernetes configs. Cloud Run handles HTTPS, scaling, and infrastructure automatically. Two million requests free per month.

Start Free

Rate This Project

Login To Rate This Project

User Reviews

Be the first to post a review of director-ai!

Additional Project Details

Registered

2026-03-29

Report inappropriate content

Ship Agents Faster

Transform your applications and workflows into powerful agentic systems at global scale.

Gemini Enterprise Agent Platform lets you rapidly build, scale, govern and optimize production-ready agents grounded in your organization's data. The platform enables developers to build custom or pre-built agents for virtually any use case. New customers get $300 in free credits.

Start Free

Recommended Projects

Active Director
The Active Director offers an extensible administration interface for servers and workstations in LDAP, WinNT, IIS and NDS environments by combining the powers of ADSI and WMI technologies. Mass updates, import/export, automation and many more await you.
DeepEval
DeepEval is a simple-to-use, open-source LLM evaluation framework, for evaluating and testing large-language model systems. It is similar to Pytest but specialized for unit testing LLM outputs. DeepEval incorporates the latest research to evaluate LLM outputs based on metrics such as G-Eval, hallucination, answer relevancy, RAGAS, etc., which uses LLMs and various other NLP models that run locally on your machine for evaluation. Whether your application is implemented via RAG or fine-tuning, LangChain, or LlamaIndex, DeepEval has you covered. With it, you can easily determine the optimal hyperparameters to improve your RAG pipeline, prevent prompt drifting, or even transition from OpenAI to hosting your own Llama2 with confidence.
OpenSMUS
OpenSMUS is a multiuser server application implementing the Shockwave MultiUser protocol published by Macromedia and Adobe, to be used with Director movies and Shockwave games.
ASSERT
Requirement-driven evaluation harness for AI agents and LLM
Intel(R) RDT Software Package
User space software for Intel(R) Resource Director Technology