Add MinimalConsole for cleaner CLI output

docs/guides/chat.mdx

@@ -72,6 +72,11 @@ gaia chat
 # Show performance metrics
 gaia chat --stats
 ```
+
+```bash Verbose Output
+# Show detailed agent output (thoughts, goals, tool results)
+gaia chat --verbose
+```
 </CodeGroup>
 
 <AccordionGroup>

docs/guides/code.mdx

@@ -171,11 +171,18 @@ your-app/
 ## Debug and Trace Options
 
 <CardGroup cols={2}>
+  <Card title="Verbose Output" icon="terminal">
+    ```bash
+    gaia-code "Create a todo tracking app in nextjs" --verbose
+    ```
+    Show detailed agent output (thoughts, goals, plans, tool results)
+  </Card>
+
   <Card title="Debug Logging" icon="bug">
     ```bash
     gaia-code "Create a todo tracking app in nextjs" --debug
     ```
-    See internal decision logs
+    See internal decision logs (implies --verbose)
   </Card>
 
   <Card title="JSON Trace" icon="file-code">

docs/reference/cli.mdx

@@ -1222,9 +1222,14 @@ All commands support these global options:
 
 | Option | Type | Default | Description |
 |--------|------|---------|-------------|
+| `--verbose` | flag | off | Show detailed agent output (thoughts, goals, plans, tool results). Default is minimal, clean output. |
 | `--logging-level` | string | INFO | Logging verbosity [DEBUG, INFO, WARNING, ERROR, CRITICAL] |
 | `-v, --version` | flag | - | Show program's version and exit |
 
+<Note>
+By default, GAIA uses a **minimal console** that shows only LLM answers, brief tool status with spinners, and errors. Use `--verbose` to see the full step-by-step agent output including thoughts, goals, plans, and detailed tool results.
+</Note>
+
 ---
 
 ## Troubleshooting

docs/sdk/core/agent-system.mdx

@@ -365,6 +365,7 @@ agent = MyAgent(
     max_steps=20,                  # Max reasoning loop iterations
     streaming=True,                # Stream responses token-by-token
     silent_mode=False,             # Suppress console output
+    verbose=False,                 # Use full AgentConsole output (default: MinimalConsole)
 
     # === Debugging ===
     debug_prompts=False,           # Print raw prompts to console
@@ -464,7 +465,8 @@ agent = MyAgent(
 |-----------|-----------|------------|-----------|
 | `max_steps` | `3` - Quick tasks | `20` - Complex workflows | Speed vs. thoroughness |
 | `streaming` | `False` - Wait for complete response | `True` - See tokens as generated | Latency vs. responsiveness |
-| `silent_mode` | `False` - See all output | `True` - Only get results | Visibility vs. clean output |
+| `verbose` | `False` - Clean, minimal output (default) | `True` - Full step-by-step agent output | Clarity vs. visibility |
+| `silent_mode` | `False` - See output | `True` - Only get results | Visibility vs. programmatic use |
 
 ---
 

docs/sdk/core/console.mdx

@@ -8,7 +8,7 @@ icon: "terminal"
 </Info>
 
 <Note>
-**Import:** `from gaia.agents.base.console import AgentConsole, SilentConsole`
+**Import:** `from gaia.agents.base.console import MinimalConsole, AgentConsole, SilentConsole`
 </Note>
 ---
 
@@ -20,23 +20,54 @@ icon: "terminal"
 
 ---
 
-## AgentConsole (Rich CLI Output)
+## MinimalConsole (Default)
+
+```python
+from gaia.agents.base.console import MinimalConsole
+
+class MyAgent(Agent):
+    def _create_console(self):
+        # Clean, minimal output (default behavior)
+        return MinimalConsole()
+
+# Shows:
+# - LLM answers (streamed as markdown)
+# - Tool calls (one-line with spinner)
+# - Brief tool output (key results, not raw JSON)
+# - Errors and warnings (inline)
+#
+# Suppresses:
+# - Agent internals (thoughts, goals, plans, step numbers)
+# - Full JSON dumps of tool arguments/results
+# - Decorative panels and separators
+```
+
+`MinimalConsole` is the **default** output handler for all agents. It provides a clean, fast user experience by showing only what matters: answers, brief tool status, and errors.
+
+Use `--verbose` on the CLI (or `verbose=True` in code) to switch to `AgentConsole`.
+
+---
+
+## AgentConsole (Verbose CLI Output)
 
 ```python
 from gaia.agents.base.console import AgentConsole
 
 class MyAgent(Agent):
     def _create_console(self):
-        # Rich console with syntax highlighting
+        # Rich console with full step-by-step output
         return AgentConsole()
 
 # Displays:
 # - Colored step headers
-# - Syntax-highlighted tool calls
+# - Thoughts, goals, and plans
+# - Syntax-highlighted tool calls with full JSON
 # - Progress indicators
-# - Formatted results
+# - Formatted results with panels
 ```
 
+`AgentConsole` is used when `--verbose` is passed on the CLI or `verbose=True` in the agent constructor. It provides maximum visibility into the agent's reasoning process, useful for development and debugging.
+
 ---
 
 ## SilentConsole (No Output)
@@ -77,7 +108,8 @@ agent = MyAgent(output_handler=handler)
 
 | Handler | Use Case | Output | Best For |
 |---------|----------|--------|----------|
-| `AgentConsole` | Interactive CLI | Rich, colored text | Development, debugging |
+| `MinimalConsole` | Interactive CLI (default) | Clean, minimal text | End users, production |
+| `AgentConsole` | Verbose CLI (`--verbose`) | Rich, colored text | Development, debugging |
 | `SilentConsole` | Headless operation | None | APIs, tests, background |
 | `SSEOutputHandler` | Streaming API | Server-Sent Events | Web applications |
 

docs/spec/console.mdx

@@ -7,9 +7,9 @@ title: "Console Output Handlers"
 </Info>
 
 <Note>
-**Component:** OutputHandler, AgentConsole, SilentConsole, SSEOutputHandler
+**Component:** OutputHandler, MinimalConsole, AgentConsole, SilentConsole, SSEOutputHandler
 **Module:** `gaia.agents.base.console`, `gaia.api.sse_handler`
-**Import:** `from gaia.agents.base.console import OutputHandler, AgentConsole, SilentConsole`
+**Import:** `from gaia.agents.base.console import OutputHandler, MinimalConsole, AgentConsole, SilentConsole`
 </Note>
 ---
 
@@ -32,22 +32,25 @@ The console output system provides a unified interface for handling agent output
 ```mermaid
 %%{init: {'theme':'base', 'themeVariables': { 'primaryColor':'#ED1C24', 'primaryTextColor':'#fff', 'primaryBorderColor':'#C8171E', 'lineColor':'#F4484D', 'secondaryColor':'#2d2d2d', 'tertiaryColor':'#f5f5f5', 'fontFamily': 'system-ui, -apple-system, sans-serif'}}}%%
 flowchart TD
-    A(["OUTPUT HANDLER"]) --> B(["AGENT CONSOLE"])
+    A(["OUTPUT HANDLER"]) --> B(["MINIMAL CONSOLE (default)"])
+    A --> E(["AGENT CONSOLE (--verbose)"])
     A --> C(["SILENT CONSOLE"])
     A --> D(["SSE HANDLER"])
 
     style A fill:#ED1C24,stroke:#C8171E,stroke-width:2px,color:#fff
     style B fill:#F4484D,stroke:#ED1C24,stroke-width:2px,color:#fff
+    style E fill:#F4484D,stroke:#ED1C24,stroke-width:2px,color:#fff
     style C fill:#F4484D,stroke:#ED1C24,stroke-width:2px,color:#fff
     style D fill:#F4484D,stroke:#ED1C24,stroke-width:2px,color:#fff
 
-    linkStyle 0,1,2 stroke:#ED1C24,stroke-width:2px
+    linkStyle 0,1,2,3 stroke:#ED1C24,stroke-width:2px
 ```
 
 | Component | Type | Purpose |
 |-----------|------|---------|
 | **OUTPUT HANDLER** | Abstract Base Class | Defines interface for all output handlers |
-| **AGENT CONSOLE** | Implementation | Rich CLI output with colors, spinners, syntax highlighting |
+| **MINIMAL CONSOLE** | Implementation (default) | Clean, minimal output — answers, brief tool status, errors |
+| **AGENT CONSOLE** | Implementation (`--verbose`) | Rich CLI output with colors, spinners, syntax highlighting, full step-by-step detail |
 | **SILENT CONSOLE** | Implementation | Suppressed output for testing and API mode |
 | **SSE HANDLER** | Implementation | Server-Sent Events streaming for API clients |
 

src/gaia/agents/base/agent.py

@@ -16,7 +16,7 @@
 import uuid
 from typing import Any, Dict, List, Optional
 
-from gaia.agents.base.console import AgentConsole, SilentConsole
+from gaia.agents.base.console import AgentConsole, MinimalConsole, SilentConsole
 from gaia.agents.base.errors import format_execution_trace
 from gaia.agents.base.tools import _TOOL_REGISTRY
 
@@ -76,6 +76,7 @@ def __init__(
         show_stats: bool = False,
         silent_mode: bool = False,
         debug: bool = False,
+        verbose: bool = False,
         output_handler=None,
         max_plan_iterations: int = 3,
         max_consecutive_repeats: int = 4,
@@ -99,7 +100,9 @@ def __init__(
             show_stats: If True, displays LLM performance stats after each response (default: False)
             silent_mode: If True, suppresses all console output for JSON-only usage (default: False)
             debug: If True, enables debug output for troubleshooting (default: False)
-            output_handler: Custom OutputHandler for displaying agent output (default: None, creates console based on silent_mode)
+            verbose: If True, uses the full AgentConsole with detailed step-by-step output.
+                     If False (default), uses MinimalConsole with clean, minimal output.
+            output_handler: Custom OutputHandler for displaying agent output (default: None, creates console based on mode flags)
             max_plan_iterations: Maximum number of plan-execute-replan cycles (default: 3, 0 = unlimited)
             max_consecutive_repeats: Maximum consecutive identical tool calls before stopping (default: 4)
             min_context_size: Minimum context size required for this agent (default: 32768).
@@ -120,6 +123,7 @@ def __init__(
         self.show_stats = show_stats
         self.silent_mode = silent_mode
         self.debug = debug
+        self.verbose = verbose
         self.last_result = None  # Store the most recent result
         self.max_plan_iterations = max_plan_iterations
         self.max_consecutive_repeats = max_consecutive_repeats
@@ -348,15 +352,22 @@ def _get_system_prompt(self) -> str:
     def _create_console(self):
         """
         Create and return a console output handler.
-        Returns SilentConsole if in silent_mode, otherwise AgentConsole.
+
+        Priority:
+        1. SilentConsole if silent_mode (JSON-only, no output)
+        2. AgentConsole if verbose (full step-by-step debug output)
+        3. MinimalConsole (default — clean, minimal user experience)
+
         Subclasses can override this to provide domain-specific console output.
         """
         if self.silent_mode:
             # Check if we should completely silence everything (including final answer)
             # This would be true for JSON-only output or when output_dir is set
             silence_final_answer = getattr(self, "output_dir", None) is not None
             return SilentConsole(silence_final_answer=silence_final_answer)
-        return AgentConsole()
+        if self.verbose:
+            return AgentConsole()
+        return MinimalConsole()
 
     @abc.abstractmethod
     def _register_tools(self):