Privacy and governance in the AI bot era are front and center. Bot traffic is real: hundreds of AI bot visits daily on LLMS Central’s network, with GPTBot accounting for about 60% of traffic and Claude and Perplexity bots also showing up. Peak crawling happens in the early hours, 2-4 AM UTC [1].
Sites with a llms.txt file see about 40% more organized crawling, and the setup touts real-time bot tracking, page-level analytics, and GDPR-compliant, non-PII data handling. The stack includes Next.js 15 and Firebase Firestore, plus a lightweight 2KB script that aims for zero perf hit while offering near 99%+ accuracy in bot detection [1].
In practice, a Reddit thread about building a private coaching/therapy chatbot highlights privacy as a dealbreaker for many: people want offline, local LLMs and to avoid uploading personal data to big hosts. The thread points to offline pathways on hardware like a Mac Mini M3 Pro, with starting points including llama.cpp and GUI options such as Jan Desktop or LM Studio—and notes that “Apple silicon is a first-class citizen” for these setups [2].
Practical takeaways: - For developers: offer on-device processing, transparent data retention controls, and clear permissions; consider adopting a llms.txt-style approach to signal which AI systems may cite your content [1][2]. - For users: prefer privacy-preserving, local options when possible; verify hardware readiness (e.g., Mac Mini M3 Pro) and explore tools like llama.cpp, Jan Desktop, and LM Studio for personal coaching apps [2].
Takeaway: balance cloud convenience with on-device safety, and watch how standards like llms.txt and local tooling shape privacy in 2025 [1][2].
References
Show HN: We're tracking AI bot visits daily across our network
Show HN: LLMS Central tracks daily AI bot visits, analyzes crawl patterns, privacy, llms.txt proposal, and questions bot permissions.
View sourceTotal noob here who wants to run a local LLM to build my own coach and therapist chatbot
Begins with local LLMs; compares tools (llama.cpp, Koboldcpp, Open WebUI) and cloud options; warns therapy risks privacy, context, cost concerns.
View source