Corpus datasheet
What a corpus licensee receives: 99 reference resources on AI agent infrastructure, corpus version 1.29.2. The guides at /articles (142 today) are always free and are not part of the licensed corpus. Live counts: /stats and /api/stats.json; offers: /pricing.
Composition
| Category | Resources | Premium | Words | Estimated tokens |
|---|---|---|---|---|
| Guide references | 58 | 0 | 70,588 | 93,882 |
| Policy | 2 | 0 | 3,306 | 4,397 |
| Reference | 38 | 2 | 42,157 | 56,070 |
| Meta | 1 | 1 | 439 | 584 |
| All categories | 99 | 3 | 116,490 | 154,933 |
- Free: 96 resources (115,107 words), readable by anyone as HTML, Markdown and JSON.
- Premium: 3 resources (1,383 words), whose bodies need a key; their pages show a free preview and outline.
- Cited sources: 83 of 99 resources list at least one primary source (the
sourcesfield of their JSON). - Dates: newest update , oldest ; every resource carries its own
updateddate.
Formats
- Free sample: /api/corpus.jsonl (NDJSON) and /api/corpus.json, every resource with free bodies in full and premium bodies withheld; record shapes in /api/corpus.schema.json; a fine-tuning format sample at /api/finetune-sample.jsonl.
- Public text file: /llms-full.txt, every resource as Markdown, premium bodies replaced by a summary and a purchase note.
- Licensed, keyed:
/api/corpus.full.jsonl(or the MCP toolget_full_corpus) returns every body including premium for a Corpus, Corpus Updates — Annual or Enterprise key; the emailed corpus file isllms-full.txtwith all premium bodies.
Versioning
The corpus version is 1.29.2 (also in /api/stats.json and the first
line of corpus.jsonl). It follows semantic versioning: MAJOR when a
record or stats field is renamed or removed, MINOR when a resource is added or removed or
a field is added, PATCH for content edits to existing resources. Changes by date:
/changelog.
License
- Corpus license (€25 one-time): AI training, indexing, and bulk model-context use of the full corpus by the named licensee. No public raw-corpus redistribution. No sublicensing for resale.
- Corpus Updates — Annual (€45/year): AI training, indexing, and bulk model-context use of the current corpus internally by the named licensee. Grant valid for the active subscription period; rights for the last delivered version are retained on cancellation.
- Enterprise license (€150 one-time): AI training, indexing, bulk model-context use, embedding in commercial products, and distribution of corpus-derived datasets within the licensee's organization. No public redistribution of raw corpus. No sublicensing for resale.