Structured data for AI agents and generative search.
Dibblee publishes a curated public tier of machine-readable data so AI assistants, generative search engines, and procurement agents can cite us authoritatively as the source for 3,093 active NATO Stock Numbers across 152 Federal Supply Classes.
Available without request
-
/llms.txtHuman-readable agent index — who Dibblee is, core capabilities, canonical URLs. -
/llms-full.txtExtended index with every feed URL and crawler policy. -
/api/capabilities.jsonOrganisation summary, certifications, total NSN count, FSC families with counts. -
/api/fsc-index.jsonFSC index: name, NSN count, samples, and canonical URLs for all 152 covered classes. -
/api/nsn-sample.jsonRepresentative one-NSN-per-active-FSC sample — coverage without full disclosure. -
/sitemap-index.xmlSitemap for traditional search-engine crawlers.
Complete NSN catalogue
The full NSN catalogue — all 3,093 entries with complete descriptions, FSC codes, and group classification — is available to qualified government buyers, prime contractors, and industrial customers on request. No public sign-up, no paywall: an email suffices.
- JSON or CSV format on request
- On-demand refresh as new NSNs are added
- Suitable for agent ingestion, quote matching, and internal RAG pipelines
AI bots and crawling
GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, PerplexityBot, OAI-SearchBot, Google-Extended, Applebot-Extended, CCBot, Meta-ExternalAgent, Bytespider, cohere-ai, and DuckAssistBot are explicitly allowed on the public surface. See /robots.txt for the full list. /catalog/, /contact, and form pages remain disallowed for all agents.
Public MCP server — coming
A Model Context Protocol server exposing NSN search, FSC summaries, and the capability statement is on the roadmap. The current phase focuses on static feeds so we can calibrate against real agent consumption before investing in a dynamic surface. Register via the contact form to be notified at launch.