#hallucination-detection

Thoth — agentic systematic literature reviews with per-claim citation verification (cite_check). 8-stage LangGraph agent, authenticated MCP server on the official registry, public eval dashboard, Langfuse/OTel tracing. 676 tests, $0/mo.

Trust layer for document→JSON extraction & AI agents: calibrated per-field confidence + source grounding + accept/review abstention on any OCR/VLM. Ships VerifyDocBench, a novel grounding-conditioned conformal method, and an MCP server.

Trustworthy AI evaluation tool for testing chatbot safety, reliability, hallucination behavior, privacy risk, and instruction-following quality.

Real-time detection and repair of LLM agent failures — a one-class behavioural monitor at ~200 µs/step, with 2,823 committed traces.

Backend-agnostic black-box hallucination & RAG-faithfulness detection for LLMs — Probabilistic Certainty & Consistency (arXiv:2601.02574). Works on MLX / OpenAI / vLLM via token logprobs; no model internals, no training.

RAG multi-agente con validación QA anti-alucinaciones (OpenAI + Pinecone + LangChain). Modo offline reproducible.

Fully-local terminal AI agent for Ollama. 34 tools, MCP, local RAG, deterministic hallucination checks. No cloud, no API keys, nothing leaves your machine.