feat: complete overhaul of AI subsystem with multi-agent litellm architecture

- Refactored AI module to use litellm, supporting Anthropic, Google, OpenAI, etc.
- Introduced 'Engineer' (execution) and 'Architect' (strategic) AI agents.
- Added real-time streaming responses and interactive chat mode via 'rich'.
- Added CLI arguments for model/key overrides (--engineer-model, --architect-model).
- Replaced 'openai' with 'litellm' in requirements.txt and setup.cfg.
- Updated nodes.run() to support an 'on_complete' callback for live node-status streaming.
- Fixed an undefined variable bug (config.profiles -> self.profiles) in configfile.py.
- Updated README and docstrings with new AI plugin tool registration API.
- Regenerated HTML documentation using pdoc3.
- Bumped version to 5.0b1 for beta release.

README.md

@@ -56,7 +56,9 @@ For more detailed information, please read our [Privacy Policy](https://connpy.g
       Or use fzf by installing pyfzf and running conn config --fzf true.
     - Create in bulk, copy, move, export, and import nodes for easy management.
     - Run automation scripts on network devices.
-    - Use GPT AI to help you manage your devices.
+    - Use AI with a multi-agent system (Engineer/Architect) to manage devices.
+      Supports any LLM provider via litellm (OpenAI, Anthropic, Google, etc.).
+      Features streaming responses, interactive chat, and extensible plugin tools.
     - Add plugins with your own scripts.
     - Much more!
 
@@ -428,15 +430,46 @@ for key in routers.result:
     print(key, ' ---> ', ("pass" if routers.result[key] else "fail"))
 ```
 ### Using AI
-```
+The AI module uses a multi-agent architecture with an **Engineer** (fast execution) and an **Architect** (strategic reasoning). It supports any LLM provider through [litellm](https://github.com/BerriAI/litellm).
+```python
 import connpy
 conf = connpy.configfile()
-organization = 'openai-org'
-api_key = "openai-key"
-myia = connpy.ai(conf, organization, api_key)
-input = "go to router 1 and get me the full configuration"
-result = myia.ask(input, dryrun = False)
-print(result)
+# Uses models and API keys from config, or override them:
+myai = connpy.ai(conf, engineer_model="gemini/gemini-2.5-flash", engineer_api_key="your-key")
+result = myai.ask("go to router1 and show me the running configuration")
+print(result["response"])
+# Streaming is enabled by default for CLI, disable for programmatic use:
+result = myai.ask("show interfaces on all routers", stream=False)
+print(result["response"])
+```
+
+#### AI Plugin Tool Registration
+Plugins can extend the AI system by registering custom tools via the `Preload` class:
+```python
+def _register_my_tools(ai_instance):
+    tool_def = {
+        "type": "function",
+        "function": {
+            "name": "my_custom_tool",
+            "description": "Does something useful.",
+            "parameters": {
+                "type": "object",
+                "properties": {"query": {"type": "string"}},
+                "required": ["query"]
+            }
+        }
+    }
+    ai_instance.register_ai_tool(
+        tool_definition=tool_def,
+        handler=my_handler_function,
+        target="engineer",  # or "architect" or "both"
+        engineer_prompt="- My tool: does X.",
+        architect_prompt="  * My tool (my_custom_tool)."
+    )
+
+class Preload:
+    def __init__(self, connapp):
+        connapp.ai.modify(_register_my_tools)
 ```
 ## http API
 With the Connpy API you can run commands on devices using http requests
@@ -527,7 +560,7 @@ With the Connpy API you can run commands on devices using http requests
 
 **Method**: `POST`
 
-**Description**: This route sends to chatgpt IA a request that will parse it into an understandable output for the application and then run the request.
+**Description**: This route sends a request to the AI multi-agent system which will analyze it, execute commands on devices if needed, and return the result. Supports any LLM provider configured via litellm.
 
 #### Request Body: