fix HowTo config

Author: fogfish

doc/HOWTO.md

@@ -2,6 +2,7 @@
 
 - [HowTos](#howtos)
   - [Setup environment](#setup-environment)
+    - [Example configurations](#example-configurations)
 
 
 ## Setup environment
@@ -16,36 +17,58 @@ Once the provider and LLMs access is configures, your application can start usin
 ```go
 import "github.com/kshard/chatter/provider/autoconfig"
 
-llm, err := autoconfig.New("thinker")
+llm, err := autoconfig.FromNetRC("thinker")
 if err != nil {
   panic(err)
 }
 ```
 
-The `autoconfig` reads the desired configuration from `~/.netrc` and creates appropriate instance of LLM API.
+The `autoconfig` reads the desired configuration from `~/.netrc` and creates appropriate instance of LLM API. Your `~/.netrc` file must include at least the `provider` and `model` fields under a named service entry. For example:
+
+```
+machine thinker
+  provider provider:bedrock/foundation/converse
+  model us.anthropic.claude-3-7-sonnet-20250219-v1:0
+```
+
+* `provider` specifies the full path to the provider's capability (e.g., `provider:bedrock/foundation/converse`). The path ressembles import path of providers implemented by this library
+* `model` specifies the exact model name as recognized by the provider
+
+Each provider and model family may support additional options. These can also be added under the same `machine` entry and will be passed into the corresponding provider implementation.
+
+```
+region     // used by Bedrock providers
+host       // used by OpenAI providers
+secret     // used by OpenAI providers
+timeout    // used by OpenAI providers
+dimensions // used by embedding families
+```
+
+### Example configurations 
+
 
 **For AWS Bedrock**, `~/.netrc` config is
 ```
 machine thinker
-  provider bedrock
+  provider provider:bedrock/foundation/converse
+  model us.anthropic.claude-3-7-sonnet-20250219-v1:0
   region us-west-2
-  family llama3
-  model meta.llama3-1-70b-instruct-v1:0
 ```
 
 **For OpenAI**, `~/.netrc` config is
 ```
 machine chatter1
-  provider openai
-  host https://api.openai.com
+  provider provider:openai/foundation/gpt
   model gpt-4o
+  host https://api.openai.com
   secret sk-...IA
 ```
 
 **For LM Studio**, `~/.netrc` config is
 ```
 machine chatter1
-  provider openai
-  host http://localhost:1234
+  provider provider:openai/foundation/gpt
   model gemma-3-27b-it
+  host http://localhost:1234
+  timeout 300
 ```