Sources
The boundary is a published list.
The owner publishes what enters. Acervo resolves that owned corpus first, then looks elsewhere only for a declared gap or needed refresh.
- scope· reading
- lenses· reading
- sources· reading
Recorded admission
Nothing is searchable by accident.
Every item travels an ordered, resumable path. A failed item stops in its own record; it does not erase the rest.
01 · discovered
Decides whether a candidate is ready for review.
states
queued · blocked · failed
02 · policy decision
Decides what may be fetched, stored, and shown.
rights states
unknown · metadata_only · licensed_storage · legal_review · denied · expired
03 · fetch and extract
Decides whether readable material can enter the corpus.
states
fetched · extracted · blocked · failed · gone
04 · digest and dedup
Decides which canonical copy the corpus retains.
job states
succeeded · partial · failed · dead_letter
05 · classify and assess
Decides which public vocabulary and evidence apply.
job states
succeeded · partial · failed · blocked
06 · embed and index
Decides whether admitted material is retrievable.
states
indexed · blocked · failed
07 · searchable
Decides whether the public directory may expose it.
freshness states
unknown · fresh · stale · refreshing · gone
Admitted sources
policy version unavailable
One row per admitted source class and rights state, with the domains and documents behind it, how much is indexed, and when it was last refreshed. Counts are measured; a blank cell is never a zero.
Rights state
What the corpus may retain
Site nature
How a source is declared
- news
- personal_blog
- forum
- store
- institutional
- aggregator
- content_farm
- reference
- other
Select a state to read its definition.
A chip carries declared vocabulary, never a computed judgement about a result and never a rank signal.
A deliberate boundary
The directory is public. Your searches are not.
shared to this deployment
- Admitted documents
- Domain assessments
- Coverage decisions
- Source policies
private to you
- Queries and search history
- Collections
- Saved searches
- Personal key usage
Coverage and freshness can be published. Shared analytics never carry query text; what you search, save, collect, and revisit stays yours.