Use host Ollama for Docker app runtime

README.md

@@ -32,17 +32,29 @@ ollama pull qwen3.5:9b
 
 ## Docker Compose
 
-Start both the app and `Ollama` together:
+Run the app in Docker while keeping `Ollama` on the host:
 
 ```bash
 docker compose up --build
 ```
 
 This compose stack will:
-- start `Ollama`
-- pull `qwen3.5:9b` through the `ollama-init` service
-- start the Next.js app on `http://localhost:3000`
-- persist the SQLite database and pulled model with named Docker volumes
+- start only the Next.js app on `http://localhost:3000`
+- keep SQLite data in a named Docker volume
+- connect to host `Ollama` through `host.docker.internal`
+
+Before running Docker Compose, make sure host `Ollama` is already up:
+
+```bash
+ollama serve
+ollama pull qwen3.5:9b
+```
+
+If you run the app outside Docker, keep using:
+
+```bash
+OLLAMA_URL=http://127.0.0.1:11434/
+```
 
 ## In-app helpers
 
@@ -53,5 +65,5 @@ This compose stack will:
 ## Environment
 
 - `DATABASE_URL` - Prisma SQLite connection string
-- `OLLAMA_URL` - local or container Ollama base URL
+- `OLLAMA_URL` - Ollama base URL; in Docker Compose this defaults to `http://host.docker.internal:11434/`
 - `OLLAMA_MODEL` - selected model tag, default `qwen3.5:9b`