.env.example

# ── LLM provider ─────────────────────────────────────────────────────────────
# Which backend Plume calls:
#   anthropic (default) — native Anthropic Messages API.
#   openai              — ANY OpenAI-compatible /chat/completions endpoint:
#                         OpenAI, OpenRouter, Groq, Together, DeepSeek, Mistral,
#                         Google's OpenAI-compat layer, or a local runtime like
#                         Ollama / LM Studio / vLLM.
LLM_PROVIDER=anthropic

# ── Anthropic  (used when LLM_PROVIDER=anthropic) ────────────────────────────
# Get a key at https://console.anthropic.com/settings/keys
ANTHROPIC_API_KEY=sk-ant-...

# ── OpenAI-compatible  (used when LLM_PROVIDER=openai) ───────────────────────
# Base URL of the API, up to the version segment (no /chat/completions):
#   OpenAI      https://api.openai.com/v1
#   OpenRouter  https://openrouter.ai/api/v1
#   Groq        https://api.groq.com/openai/v1
#   Ollama      http://localhost:11434/v1   (local — no key needed)
#   LM Studio   http://localhost:1234/v1    (local — no key needed)
# LLM_BASE_URL=https://api.openai.com/v1
# LLM_API_KEY=sk-...

# ── Model ────────────────────────────────────────────────────────────────────
# Model id to use. Defaults to claude-sonnet-4-6 for the anthropic provider
# (cheaper: claude-haiku-4-5). REQUIRED and provider-specific for openai, e.g.:
#   OpenAI  gpt-4o-mini     OpenRouter  anthropic/claude-3.5-sonnet
#   Groq    llama-3.3-70b-versatile      Ollama  llama3.1
MODEL=claude-sonnet-4-6

# ── Server ───────────────────────────────────────────────────────────────────
# Port the Express API listens on (default 3001).
PORT=3001

# ── Rate limiting (protects your key / endpoint on public deploys) ───────────
# Max requests per IP per window, and window length in milliseconds.
RATE_LIMIT_MAX=30
RATE_LIMIT_WINDOW_MS=300000

# Set to 1 (or the number of proxy hops) ONLY if deployed behind a reverse
# proxy / load balancer, so rate limiting sees the real client IP.
# TRUST_PROXY=1