Show HN: Synthetic corporate dataset generator for AI agent evaluation

Category: ai-ml

Tags: synthetic-data, ai-evaluation, enterprise-simulation, data-generation, llm-testing

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 6)

OrgForge generates realistic synthetic enterprise datasets (emails, JIRA tickets, Slack messages, etc.) via a deterministic state machine, enabling evaluation of AI agents on grounded, multi-source corporate knowledge. Its key innovation is coupling LLM-generated prose with a controlled event log, preventing hallucination and providing ground truth for testing retrieval and reasoning.

Target audience: data engineers, ai researchers, devops, backend devs

Repository: https://github.com/aeriesec/orgforge · Python · MIT · 13 stars

View on Hacker News