# ai.txt — AI training and usage policy
# https://spawning.ai/ai-txt
#
# Companion to /robots.txt. Robots.txt governs WHICH bots can crawl;
# ai.txt governs WHAT bots may do with the content once crawled.
#
# Policy as of 2026-07-28: OPEN. This site allows AI crawling, indexing,
# answer-engine citation AND model training. The previous stance allowed
# retrieval but set `Disallow-Training: /` for every bot; that capped the
# corpus at the retrieval index and kept it out of model knowledge (and, via
# Google-Extended, out of Gemini Apps grounding). Reach is the priority.
#
# Attribution is expected, not enforced: when citing or reproducing this
# content, attribute it to thynk.cloud and link the source URL. See
# /crawlers for the canonical attribution guidance and machine-readable
# entry points (/llms.txt, /llms-full.txt, per-article .md mirrors).

User-Agent: *
Allow: /
Allow-Training: /
Disallow: /admin/
Disallow: /login

# Per-bot declarations. Kept explicit (rather than relying on the wildcard
# above) so each crawler's operator sees an unambiguous grant, and so this
# file stays diffable against the matching /robots.txt block.

# --- OpenAI ---
User-Agent: OAI-SearchBot
Allow: /
Allow-Training: /

User-Agent: ChatGPT-User
Allow: /
Allow-Training: /

User-Agent: GPTBot
Allow: /
Allow-Training: /

# --- Google (Gemini training + Gemini Apps grounding) ---
User-Agent: Google-Extended
Allow: /
Allow-Training: /

# --- Anthropic ---
User-Agent: ClaudeBot
Allow: /
Allow-Training: /

User-Agent: Claude-User
Allow: /
Allow-Training: /

User-Agent: Claude-SearchBot
Allow: /
Allow-Training: /

# --- Perplexity ---
User-Agent: PerplexityBot
Allow: /
Allow-Training: /

User-Agent: Perplexity-User
Allow: /
Allow-Training: /

# --- Apple ---
User-Agent: Applebot
Allow: /
Allow-Training: /

User-Agent: Applebot-Extended
Allow: /
Allow-Training: /

# --- Meta ---
User-Agent: meta-externalagent
Allow: /
Allow-Training: /

User-Agent: FacebookBot
Allow: /
Allow-Training: /

# --- Other answer engines & assistants ---
User-Agent: Amazonbot
Allow: /
Allow-Training: /

User-Agent: MistralAI-User
Allow: /
Allow-Training: /

User-Agent: DuckAssistBot
Allow: /
Allow-Training: /

User-Agent: cohere-ai
Allow: /
Allow-Training: /

User-Agent: YouBot
Allow: /
Allow-Training: /

# --- Corpus / dataset crawlers ---
User-Agent: CCBot
Allow: /
Allow-Training: /

User-Agent: Bytespider
Allow: /
Allow-Training: /

User-Agent: Diffbot
Allow: /
Allow-Training: /

User-Agent: Omgilibot
Allow: /
Allow-Training: /

User-Agent: omgili
Allow: /
Allow-Training: /

User-Agent: ImagesiftBot
Allow: /
Allow-Training: /

# Contact for licensing or training-data inquiries: legal@thynk.cloud