diff --git a/docs/configuration.md b/docs/configuration.md
index dadf6015..c03640af 100644
--- a/docs/configuration.md
+++ b/docs/configuration.md
@@ -129,9 +129,11 @@ This is useful for:
 - Testing with different configurations
 - Applications that need multiple clients with different configurations
 
-## API Keys
+## Environment Variables
 
-API keys are configured through **environment variables**, not in the YAML file.
+API keys and some provider settings are configured through **environment variables**, not in the YAML file.
+
+### API Keys
 
 ```bash
 # OpenAI
@@ -147,6 +149,13 @@ export VOYAGE_API_KEY=your-key-here
 export CO_API_KEY=your-key-here
 ```
 
+### Provider Configuration
+
+```bash
+# Ollama base URL (defaults to http://localhost:11434)
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
 ## File Monitoring
 
 Set directories to monitor for automatic indexing:
@@ -171,6 +180,23 @@ embeddings:
   vector_dim: 1024
 ```
 
+The Ollama base URL can be configured via environment variable or config file:
+
+```bash
+# Via environment variable (recommended)
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
+Or in your config file:
+
+```yaml
+providers:
+  ollama:
+    base_url: http://localhost:11434
+```
+
+If neither is set, it defaults to `http://localhost:11434`.
+
 ### VoyageAI
 
 If you want to use VoyageAI embeddings you will need to install `haiku.rag` with the VoyageAI extras:
@@ -236,7 +262,17 @@ Configure which LLM provider to use for question answering. Any provider and mod
 qa:
   provider: ollama
   model: gpt-oss
+```
 
+The Ollama base URL can be configured via the `OLLAMA_BASE_URL` environment variable, config file, or defaults to `http://localhost:11434`:
+
+```bash
+export OLLAMA_BASE_URL=http://localhost:11434
+```
+
+Or in your config file:
+
+```yaml
 providers:
   ollama:
     base_url: http://localhost:11434
diff --git a/src/haiku/rag/config/models.py b/src/haiku/rag/config/models.py
index 3bcc59d1..faa70c84 100644
--- a/src/haiku/rag/config/models.py
+++ b/src/haiku/rag/config/models.py
@@ -46,7 +46,11 @@ class ProcessingConfig(BaseModel):
 
 
 class OllamaConfig(BaseModel):
-    base_url: str = "http://localhost:11434"
+    base_url: str = Field(
+        default_factory=lambda: __import__("os").environ.get(
+            "OLLAMA_BASE_URL", "http://localhost:11434"
+        )
+    )
 
 
 class VLLMConfig(BaseModel):