GET /xrpc/app.bsky.actor.searchActorsTypeahead
typeahead.waow.tech
fix architecture.md: update moderation policy
!no-unauthenticated is not filtered — it applies to content, not
identity. only bluesky mod service labels (!hide, !takedown, spam)
hide actors. also removed stale ingester description about detecting
!no-unauthenticated self-labels.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+9
-9
+9
-9
docs/architecture.md
+9
-9
docs/architecture.md
···
23
23
24
24
1. **ingester**: streams jetstream, buffers actor events, POSTs batches to
25
25
`/admin/ingest`. sends avatar CIDs (not full URLs — the CDN URL is
26
-
deterministic from DID + CID). detects `!no-unauthenticated` self-labels
27
-
and sets `hidden` flag.
26
+
deterministic from DID + CID).
28
27
29
28
2. **backfill**: when a search has gaps (missing avatars, few results), the
30
29
worker calls bluesky's `searchActorsTypeahead` API and upserts results.
···
64
63
65
64
## moderation
66
65
67
-
actors with certain labels are hidden from search results (`hidden = 1`):
66
+
actors with certain labels from bluesky's moderation service
67
+
(`did:plc:ar7c4by46qjdydhdevvrndac`) are hidden from search results
68
+
(`hidden = 1`): `!hide`, `!takedown`, or `spam`.
68
69
69
-
- bluesky moderation service (`did:plc:ar7c4by46qjdydhdevvrndac`) issued
70
-
`!hide`, `!takedown`, or `spam`
71
-
- any source (including self-labels): `!no-unauthenticated`
70
+
`!no-unauthenticated` is intentionally **not** filtered — it applies to
71
+
content, not identity. bluesky's own public typeahead API returns these
72
+
accounts, and so do we.
72
73
73
-
this is an unauthenticated service, so we respect user opt-outs. the ingester
74
-
catches self-labels on ingest; the hourly cron refreshes moderation labels
75
-
by walking the full index over multiple runs.
74
+
the hourly cron refreshes moderation labels by walking the full index over
75
+
multiple runs (1000 actors/run via `mod_cursor` in KV).
76
76
77
77
## scripts
78
78