FinSight / .env.example
Sanjam19's picture
Deploy FinSight demo (single-container Docker Space)
d4f8959 verified
Raw
History Blame Contribute Delete
662 Bytes
# Copy to .env and fill in what you use.
# Cloud LLM (optional — only needed if LLM_BACKEND=groq)
GROQ_API_KEY=your_groq_api_key_here
# LLM backend selection: "ollama" (default, fully local) or "groq" (cloud, faster)
LLM_BACKEND=ollama
# Local model served by Ollama (any instruct model you have pulled).
# Small default so it runs in ~2GB RAM; use a bigger model if you have memory.
OLLAMA_MODEL=qwen2.5:1.5b
# Context window cap — keeps Ollama's memory budget small
OLLAMA_NUM_CTX=4096
# Run the local-LLM pass for ambiguous entity relations during /upload.
# Off by default: it costs one model call per ambiguous org pair.
RELATION_LLM_FALLBACK=false