enable token usage by default (#2841)

backend/packages/harness/deerflow/config/token_usage_config.py

@@ -4,4 +4,4 @@ from pydantic import BaseModel, Field
 class TokenUsageConfig(BaseModel):
     """Configuration for token usage tracking."""
 
-    enabled: bool = Field(default=False, description="Enable token usage tracking middleware")
+    enabled: bool = Field(default=True, description="Enable token usage tracking middleware")

backend/tests/test_setup_wizard.py

@@ -313,7 +313,7 @@ class TestWriteConfigYaml:
                 {
                     "config_version": 5,
                     "log_level": "info",
-                    "token_usage": {"enabled": False},
+                    "token_usage": {"enabled": True},
                     "tool_groups": [{"name": "web"}, {"name": "file:read"}, {"name": "file:write"}, {"name": "bash"}],
                     "tools": [
                         {
@@ -361,7 +361,7 @@ class TestWriteConfigYaml:
             data = yaml.safe_load(f)
 
         assert data["log_level"] == "info"
-        assert data["token_usage"]["enabled"] is False
+        assert data["token_usage"]["enabled"] is True
         assert data["tool_groups"][0]["name"] == "web"
         assert data["summarization"]["max_tokens"] == 2048
         assert any(tool["name"] == "image_search" and tool["max_results"] == 5 for tool in data["tools"])

backend/tests/test_token_usage_config.py

@@ -0,0 +1,5 @@
+from deerflow.config.token_usage_config import TokenUsageConfig
+
+
+def test_token_usage_enabled_by_default():
+    assert TokenUsageConfig().enabled is True

config.example.yaml

@@ -30,7 +30,7 @@ log_level: info
 # When enabled, DeerFlow records input/output/total tokens per model call
 # and shows usage metadata in the workspace UI when providers return it.
 token_usage:
-  enabled: false
+  enabled: true
 
 # ============================================================================
 # Models Configuration

frontend/src/content/en/harness/middlewares.mdx

@@ -110,7 +110,7 @@ Tracks LLM token consumption per model call and logs it at the `info` level. Use
 
 ```yaml
 token_usage:
-  enabled: false
+  enabled: true
 ```
 
 ---

frontend/src/content/zh/harness/middlewares.mdx

@@ -110,7 +110,7 @@ title:
 
 ```yaml
 token_usage:
-  enabled: false
+  enabled: true
 ```
 
 ---

frontend/tests/unit/core/settings/local.test.ts

@@ -0,0 +1,10 @@
+import { expect, test } from "vitest";
+
+import { DEFAULT_LOCAL_SETTINGS } from "@/core/settings/local";
+
+test("defaults token usage to header total plus per-turn breakdown", () => {
+  expect(DEFAULT_LOCAL_SETTINGS.tokenUsage).toEqual({
+    headerTotal: true,
+    inlineMode: "per_turn",
+  });
+});