diff --git a/cluster/applications/hindsight/values.yaml b/cluster/applications/hindsight/values.yaml index ee74b5f..6abd024 100644 --- a/cluster/applications/hindsight/values.yaml +++ b/cluster/applications/hindsight/values.yaml @@ -49,19 +49,22 @@ existingSecret: hindsight-credentials # ---------------------------------------------------------------------------- # API container environment (explicit env entries; the chart renders this map -# to individual env vars). LLM points at the LOCAL astro-orbiter Qwen node -# (OpenAI-compatible) — moved OFF the Nous free tier (solar-pro4 was 404/dead). +# to individual env vars). LLM points at the Nous free-tier inference endpoint +# (https://inference-api.nousresearch.com/v1), model upstage/solar-pro4:free +# (tool-calling, verified retain/recall). stepfun/step-3.7-flash was swapped out +# because it rejects Hindsight's tagged fact-extraction (BadRequestError 400 +# 'missing tags'); solar-pro4 is the previously-verified-good Nous model for the +# extract/retain path (tasks t_e3375410 / t_d0dffc3d). # HINDSIGHT_API_LLM_API_KEY is NOT set here — it comes from the existingSecret -# via envFrom (1Password `nous` item); the local node ignores it but the secret -# still satisfies the chart's envFrom contract. +# via envFrom (1Password `nous` item). # ---------------------------------------------------------------------------- api: env: HINDSIGHT_API_LLM_BASE_URL: "https://inference-api.nousresearch.com/v1" HINDSIGHT_API_LLM_PROVIDER: "openai" - # Interim swap (2026-08-28, Ryan directive): local astro-orbiter Qwen is DOWN, - # so back Hindsight with Nous free stepfun/step-3.7-flash until the local node recovers. - HINDSIGHT_API_LLM_MODEL: "stepfun/step-3.7-flash:free" + # Swap (2026-08-28, t_8516dba2): stepfun failed retain w/ 400 'missing tags'; + # return to verified-good Nous solar-pro4:free. + HINDSIGHT_API_LLM_MODEL: "upstage/solar-pro4:free" # ---------------------------------------------------------------------------- # Ingress via the chart's native template.