p2: digest with in-sift FIFO, SQLite caches, Honcho context, recover

Dixie is capped (max_inflight 2, queue_depth 8, queue_wait 20s) and
never queued on llama-server. Streaming chat, thinking off, 45s
fallback to clean. Digest and chunk-embedding caches are
content-addressed SQLite (Honcho context is not cached). Recover
uses the guarded Fetcher. guardian-star is refused in code.
This commit is contained in:
2026-09-22 23:54:49 -07:00
parent 505e113b64
commit 2f9d3d97e6
2023 changed files with 3815943 additions and 30 deletions
+12
View File
@@ -24,3 +24,15 @@ listen = "127.0.0.1:8890"
# [digest]
# model = "ornith-1.5-9b-uncensored"
# thinking = false
# cache = "/var/lib/sift/cache.db"
#
# [dixie]
# chat = "http://dixie.scylla-hammerhead.ts.net:11434/v1/chat/completions"
# embed = "http://straylight.scylla-hammerhead.ts.net:11437/v1/embeddings"
# max_inflight = 2
# queue_depth = 8
# queue_wait = "20s"
#
# [context]
# honcho = "http://rift.scylla-hammerhead.ts.net:8000"
# workspaces = ["hermes-fleet"]