env.example: document LLAMA_SWAP_* + HYBRID_VISION_BACKEND vars

Mirrors the section added to CLAUDE.md so deploys can opt into the
llamacpp backend from the template alone.

.env.example

@@ -53,6 +53,27 @@ AGENTIC_CHAT_MAX_ITERATIONS=6
 # OPENROUTER_HTTP_REFERER=https://your-site.example
 # OPENROUTER_APP_TITLE=ImageApi
 
+# ── AI Insights — llama.cpp / llama-swap (optional) ─────────────────────
+# Set LLAMA_SWAP_URL to enable the `llamacpp` chat_backend. Talks
+# OpenAI-compatible /v1 to a llama-swap proxy that fronts per-slot
+# llama-server instances (chat / vision / embed). Like hybrid, the
+# agentic loop describes images via the vision slot then inlines the
+# text into the chat slot — so the chat slot itself can be text-only.
+# LLAMA_SWAP_URL=http://localhost:9292/v1
+# LLAMA_SWAP_PRIMARY_MODEL=chat
+# LLAMA_SWAP_VISION_MODEL=vision
+# LLAMA_SWAP_EMBEDDING_MODEL=embed
+# Comma-separated allowlist of model ids the /v1/models endpoint should
+# advertise as vision-capable (llama-swap doesn't report modality).
+# LLAMA_SWAP_VISION_MODELS=vision
+# Comma-separated allowlist surfaced by /insights/llamacpp/models.
+# LLAMA_SWAP_ALLOWED_MODELS=chat,vision,embed
+# LLAMA_SWAP_REQUEST_TIMEOUT_SECONDS=120
+# Routes hybrid mode's vision-describe pass through llama-swap's vision
+# slot instead of Ollama (chat still goes to OpenRouter). Values:
+# `ollama` (default) | `llamacpp`.
+# HYBRID_VISION_BACKEND=ollama
+
 # ── AI Insights — sibling services (optional) ───────────────────────────
 # Apollo (places, face inference, CLIP encoders). Single-Apollo deploys
 # typically set only APOLLO_API_BASE_URL and let the face + CLIP