# ai.txt - AI usage declaration (well-known URI)
# Format: draft-car-ai-txt-wellknown-00
# https://datatracker.ietf.org/doc/draft-car-ai-txt-wellknown/
#
# APISphere publishes pay-per-call APIs for AI agents. Public docs,
# discovery files and metadata are free to crawl, index, quote and train on.
# Crawl allowlist:  https://apisphere.us.ci/ai-bot.txt
# Crawl policy:     https://apisphere.us.ci/robots.txt
# Documentation:    https://apisphere.us.ci/llms.txt and https://apisphere.us.ci/openapi.json

Spec-Version: 1.0

Site-Name: APISphere
Site-URL: https://apisphere.us.ci
Policy-URL: https://apisphere.us.ci/llms.txt

# Public pages: crawling, indexing, caching and training are all allowed.
Training: allow
Scraping: allow
Indexing: allow
Caching: allow

# Same license as the source code: MIT.
Training-License: MIT

# Rate limit for crawlers of the docs. The paid product endpoints under
# /api/v1/* are metered and paid per call via x402 (USDC on Base).
Agent: *
  Rate-Limit: 60/minute

# model training - Collects content that may be used to train OpenAI models.
Agent: GPTBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Builds the results and citations shown in ChatGPT search.
Agent: OAI-SearchBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# user-triggered answers - User-triggered page fetch inside ChatGPT when someone asks.
Agent: ChatGPT-User
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Collects content that may be used to train Claude models.
Agent: ClaudeBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Builds the search results cited inside Claude.
Agent: Claude-SearchBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# user-triggered answers - User-triggered page fetch inside Claude when someone asks.
Agent: Claude-User
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Controls Gemini training and grounding.
Agent: Google-Extended
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Google Search index, including AI Overviews and AI Mode.
Agent: Googlebot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Bing index, which also grounds Copilot answers.
Agent: Bingbot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Apple search surfaces (Siri, Spotlight).
Agent: Applebot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Controls Apple Intelligence training and grounding.
Agent: Applebot-Extended
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Builds the index behind Perplexity answers.
Agent: PerplexityBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# user-triggered answers - User-triggered fetch from Perplexity.
Agent: Perplexity-User
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Powers Amazon assistant surfaces.
Agent: Amazonbot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Meta AI training and indexing.
Agent: meta-externalagent
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Link previews and other Meta surfaces.
Agent: FacebookBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - ByteDance / TikTok AI training.
Agent: Bytespider
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Collects content that may train DeepSeek models.
Agent: DeepSeekBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Training and retrieval for Cohere systems.
Agent: cohere-ai
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# user-triggered answers - User-triggered page fetch in Le Chat.
Agent: MistralAI-User
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# user-triggered answers - Fetches pages to answer questions in DuckDuckGo.
Agent: DuckAssistBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Builds the index behind You.com search.
Agent: YouBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Extracts pages into a knowledge graph.
Agent: Diffbot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Collects training data for Timpi search.
Agent: Timpibot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# model training - Collects image data from public pages.
Agent: ImagesiftBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# search index and citations - Feeds Webz.io web-data products.
Agent: Omgilibot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# open research dataset - Open crawl archive reused as training data by many labs.
Agent: CCBot
  Training: allow
  Scraping: allow
  Rate-Limit: 60/minute

# Content requirements
# Attribution: keep the APISphere name and a link when you republish.
Attribution: required
# AI-Disclosure: say when a model produced an answer that used this page.
AI-Disclosure: required