diff --git a/servers/hindsight/hindsight/.env.example b/servers/hindsight/hindsight/.env.example index da3049a..6be2ce6 100644 --- a/servers/hindsight/hindsight/.env.example +++ b/servers/hindsight/hindsight/.env.example @@ -70,15 +70,24 @@ HINDSIGHT_API_EMBEDDINGS_OPENROUTER_MODEL=baai/bge-m3 HINDSIGHT_API_EMBEDDINGS_OPENAI_BATCH_SIZE=32 HINDSIGHT_API_EMBEDDINGS_MAX_INPUT_TOKENS=8192 -# Local TEI reranker, then OpenRouter, then fail-open RRF. -HINDSIGHT_API_RERANKER_PROVIDER=tei +# Google Discovery Engine via LAN LiteLLM, then TEI/OpenRouter/RRF. +HINDSIGHT_API_RERANKER_PROVIDER=litellm +HINDSIGHT_API_RERANKER_LITELLM_API_BASE=http://192.168.50.49:4000 +HINDSIGHT_API_RERANKER_LITELLM_API_KEY=replace-with-litellm-client-key +HINDSIGHT_API_RERANKER_LITELLM_MODEL=google-rerank-default +HINDSIGHT_API_RERANKER_LITELLM_TIMEOUT=10 HINDSIGHT_API_RERANKER_TEI_URL=http://192.168.50.100:18084 HINDSIGHT_API_RERANKER_TEI_BATCH_SIZE=32 HINDSIGHT_API_RERANKER_TEI_MAX_CONCURRENT=4 HINDSIGHT_API_RERANKER_TEI_HTTP_TIMEOUT=30 HINDSIGHT_API_RERANKER_MAX_CANDIDATES=100 -HINDSIGHT_API_RERANKER_1_PROVIDER=openrouter -HINDSIGHT_API_RERANKER_1_OPENROUTER_API_KEY=replace-with-openrouter-key -HINDSIGHT_API_RERANKER_1_OPENROUTER_MODEL=cohere/rerank-4-fast -HINDSIGHT_API_RERANKER_1_OPENROUTER_TIMEOUT=30 -HINDSIGHT_API_RERANKER_2_PROVIDER=rrf +HINDSIGHT_API_RERANKER_1_PROVIDER=tei +HINDSIGHT_API_RERANKER_1_TEI_URL=http://192.168.50.100:18084 +HINDSIGHT_API_RERANKER_1_TEI_BATCH_SIZE=32 +HINDSIGHT_API_RERANKER_1_TEI_MAX_CONCURRENT=4 +HINDSIGHT_API_RERANKER_1_TEI_HTTP_TIMEOUT=30 +HINDSIGHT_API_RERANKER_2_PROVIDER=openrouter +HINDSIGHT_API_RERANKER_2_OPENROUTER_API_KEY=replace-with-openrouter-key +HINDSIGHT_API_RERANKER_2_OPENROUTER_MODEL=cohere/rerank-4-fast +HINDSIGHT_API_RERANKER_2_OPENROUTER_TIMEOUT=30 +HINDSIGHT_API_RERANKER_3_PROVIDER=rrf diff --git a/servers/hindsight/hindsight/compose.yaml b/servers/hindsight/hindsight/compose.yaml index 10fb007..9fd5969 100644 --- a/servers/hindsight/hindsight/compose.yaml +++ b/servers/hindsight/hindsight/compose.yaml @@ -26,7 +26,7 @@ services: max-file: "3" hindsight: - image: ghcr.io/vectorize-io/hindsight:0.10.1-slim@sha256:4d14ce6643ecb2dacb9446e0b967a35f811ce8d0172a39131281520b4c425da9 + image: ghcr.io/vectorize-io/hindsight:0.10.3-slim@sha256:c0bd6122a1d0d62dc5094ce43150780a28a96ec70109b44c688210231a5ac33a container_name: hindsight restart: unless-stopped env_file: