Anto01 b0fde3ee60 config: default LLM extractor model haiku -> sonnet ...

Haiku was producing noisy candidates (31% accept rate on first
triage). Sonnet should give tighter extraction with fewer false
positives while still catching the same durable-fact patterns.
Override: ATOCORE_LLM_EXTRACTOR_MODEL=haiku to revert.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-12 07:31:34 -04:00

..

api

feat: expose status field on POST /memory + persist_llm_candidates script

2026-04-12 05:51:31 -04:00

context

feat: query-relevance ordering for memory selection

2026-04-11 12:47:05 -04:00

ingestion

Expand active project wave and serialize refreshes

2026-04-06 14:58:14 -04:00

interactions

feat: cleanup endpoint, auto-extraction on capture, daily cron script

2026-04-11 10:28:32 -04:00

memory

config: default LLM extractor model haiku -> sonnet

2026-04-12 07:31:34 -04:00

models

fix: schema init ordering, deploy.sh default, client BASE_URL docs

2026-04-08 19:02:57 -04:00

observability

Add Dalidou storage foundation and deployment prep

2026-04-05 18:33:52 -04:00

ops

feat: post-backup validation + retention cleanup (Tasks B & C)

2026-04-11 09:46:46 -04:00

projects

fix(P1+P2): canonicalize project names at every trust boundary

2026-04-07 08:29:33 -04:00

retrieval

feat: tunable ranking, refresh status, chroma backup + admin endpoints

2026-04-06 18:42:19 -04:00

__init__.py

deploy: version-visible /health + deploy.sh + update runbook

2026-04-08 18:08:49 -04:00

config.py

feat: tunable ranking, refresh status, chroma backup + admin endpoints

2026-04-06 18:42:19 -04:00

main.py

deploy: version-visible /health + deploy.sh + update runbook

2026-04-08 18:08:49 -04:00