fix: add stream toggle back to agent (#13155)

* add stream toggle back to agent

* chore: auto-bake note keys and regenerate backend locales/en.json [skip ci]

* ci: stop transient Windows npm/cache flakes from failing frontend tests

The nightly frontend job fails almost every day on a single Windows
Playwright shard, never on actual test logic. With fail-fast:false over
~70 Windows shards each making several cache/artifact-service calls, a
transient GitHub Actions cache/artifact hiccup on any one shard flips
needs.setup-and-test.result to failure and reds the whole nightly
(observed steps: Setup Node.js Environment, Cache Playwright Browsers,
Upload Playwright Coverage Artifact).

- Decouple npm caching from actions/setup-node (its internal cache step
  can't be made non-fatal) and replace it with a standalone
  actions/cache@v5 step marked continue-on-error.
- Mark Cache Playwright Browsers and the artifact-upload steps
  continue-on-error so an infra blip can no longer fail a test shard.

Only a real test/browser failure can fail a shard now.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* [autofix.ci] apply automated fixes

* [autofix.ci] apply automated fixes (attempt 2/3)

* [autofix.ci] apply automated fixes

---------

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

.github/workflows/typescript_test.yml

@@ -227,8 +227,22 @@ jobs:
         id: setup-node
         with:
           node-version: ${{ env.NODE_VERSION }}
-          cache: "npm"
-          cache-dependency-path: ./src/frontend/package-lock.json
+
+      # npm caching is decoupled from setup-node so a transient cache-service
+      # hiccup (frequent on Windows runners) never fails the job. See #25 / #1.
+      - name: Get npm cache directory
+        id: npm-cache-dir
+        shell: bash
+        run: echo "dir=$(npm config get cache)" >> "$GITHUB_OUTPUT"
+
+      - name: Cache npm dependencies
+        uses: actions/cache@v5
+        continue-on-error: true
+        with:
+          path: ${{ steps.npm-cache-dir.outputs.dir }}
+          key: ${{ runner.os }}-npm-${{ env.NODE_VERSION }}-${{ hashFiles('src/frontend/package-lock.json') }}
+          restore-keys: |
+            ${{ runner.os }}-npm-${{ env.NODE_VERSION }}-
 
       - name: Install Frontend Dependencies
         run: npm ci
@@ -298,8 +312,23 @@ jobs:
         id: setup-node
         with:
           node-version: ${{ env.NODE_VERSION }}
-          cache: "npm"
-          cache-dependency-path: ./src/frontend/package-lock.json
+
+      # npm caching is decoupled from setup-node so a transient cache-service
+      # hiccup (frequent on Windows runners) never fails the job. See #25 / #1.
+      - name: Get npm cache directory
+        id: npm-cache-dir
+        shell: bash
+        run: echo "dir=$(npm config get cache)" >> "$GITHUB_OUTPUT"
+
+      - name: Cache npm dependencies
+        uses: actions/cache@v5
+        continue-on-error: true
+        with:
+          path: ${{ steps.npm-cache-dir.outputs.dir }}
+          key: ${{ runner.os }}-npm-${{ env.NODE_VERSION }}-${{ hashFiles('src/frontend/package-lock.json') }}
+          restore-keys: |
+            ${{ runner.os }}-npm-${{ env.NODE_VERSION }}-
+
       - name: Install Frontend Dependencies
         run: npm ci
         working-directory: ./src/frontend
@@ -308,6 +337,7 @@ jobs:
       - name: Cache Playwright Browsers
         id: cache-playwright
         uses: actions/cache@v5
+        continue-on-error: true
         with:
           path: ${{ env.PLAYWRIGHT_BROWSERS_PATH }}
           key: playwright-${{ env.PLAYWRIGHT_VERSION }}-chromium-${{ runner.os }}
@@ -366,6 +396,7 @@ jobs:
 
       - name: Upload Test Results
         if: always()
+        continue-on-error: true
         uses: actions/upload-artifact@v6
         with:
           name: blob-report-${{ runner.os }}-${{ matrix.shardIndex }}
@@ -375,6 +406,7 @@ jobs:
 
       - name: Upload Playwright Coverage Artifact
         if: always() && hashFiles('src/frontend/coverage/playwright/**') != ''
+        continue-on-error: true
         uses: actions/upload-artifact@v6
         with:
           name: playwright-coverage-${{ matrix.shardIndex }}

src/backend/base/langflow/locales/en.json

@@ -65,6 +65,8 @@
   "components.agent.inputs.output_schema.info.3b7a7b80": "Schema Validation: Define the structure and data types for structured output. No validation if no output schema.",
   "components.agent.inputs.project_id.display_name.c7feb8db": "watsonx Project ID",
   "components.agent.inputs.project_id.info.0d91e5f5": "The project ID associated with the foundation model (IBM watsonx.ai only)",
+  "components.agent.inputs.stream.display_name.1eec3071": "Stream",
+  "components.agent.inputs.stream.info.f835b3ff": "Stream the response from the model. Streaming works only in Chat.",
   "components.agent.inputs.system_prompt.display_name.55ce85ab": "Agent Instructions",
   "components.agent.inputs.system_prompt.info.a1fa7781": "System Prompt: Initial instructions and context provided to guide the agent's behavior. Supports dynamic placeholders: {current_date}, {model_name}, {optional_user_context}.",
   "components.agent.inputs.tools.display_name.ea93d6a2": "Tools",
@@ -173,6 +175,8 @@
   "components.altkagent.inputs.project_id.info.0d91e5f5": "The project ID associated with the foundation model (IBM watsonx.ai only)",
   "components.altkagent.inputs.response_processing_size_threshold.display_name.84c35424": "Response Processing Size Threshold",
   "components.altkagent.inputs.response_processing_size_threshold.info.d6aeb95b": "Tool output is post-processed only if response exceeds this character threshold.",
+  "components.altkagent.inputs.stream.display_name.1eec3071": "Stream",
+  "components.altkagent.inputs.stream.info.f835b3ff": "Stream the response from the model. Streaming works only in Chat.",
   "components.altkagent.inputs.system_prompt.display_name.55ce85ab": "Agent Instructions",
   "components.altkagent.inputs.system_prompt.info.a1fa7781": "System Prompt: Initial instructions and context provided to guide the agent's behavior. Supports dynamic placeholders: {current_date}, {model_name}, {optional_user_context}.",
   "components.altkagent.inputs.tools.display_name.ea93d6a2": "Tools",

src/backend/tests/unit/components/models_and_agents/test_agent_component.py

@@ -451,6 +451,47 @@ class TestAgentComponent(ComponentTestBaseWithoutClient):
         # Note: The provider-specific field name mapping happens inside get_llm,
         # so we just verify max_tokens is passed correctly
 
+    async def test_should_expose_stream_input_when_agent_component_is_loaded(self, component_class, default_kwargs):
+        """Regression: the Stream toggle disappeared from the Agent after the ModelInput unification (#12025).
+
+        Given the Agent component is loaded, When its inputs are inspected,
+        Then a 'stream' input field must be present so users can control LLM streaming.
+        """
+        component = await self.component_setup(component_class, default_kwargs)
+
+        input_names = [inp.name for inp in component.inputs if hasattr(inp, "name")]
+
+        assert "stream" in input_names, "stream input field should be present on the Agent component"
+        assert hasattr(component, "stream"), "Component should have a stream attribute"
+
+    @patch("lfx.components.models_and_agents.agent.AgentComponent.get_memory_data")
+    @patch("lfx.components.models_and_agents.agent.get_llm")
+    async def test_should_pass_stream_value_to_get_llm_when_stream_input_is_enabled(
+        self, mock_get_llm, mock_get_memory_data, component_class, default_kwargs
+    ):
+        """Regression: the Agent must forward the Stream toggle value to get_llm().
+
+        Given stream=True on the Agent, When get_agent_requirements runs,
+        Then get_llm() must be called with stream=True so the LLM streams responses.
+        """
+        from unittest.mock import AsyncMock, MagicMock
+
+        mock_get_memory_data.return_value = AsyncMock(return_value=[])
+        mock_get_llm.return_value = MagicMock()
+
+        default_kwargs["stream"] = True
+        component = await self.component_setup(component_class, default_kwargs)
+
+        # validate_model_selection requires a list — set a valid model selection
+        component.model = [{"name": "gpt-4o", "provider": "OpenAI", "metadata": {}}]
+
+        await component.get_agent_requirements()
+
+        mock_get_llm.assert_called_once()
+        call_kwargs = mock_get_llm.call_args.kwargs
+        assert "stream" in call_kwargs, "stream should be passed to get_llm"
+        assert call_kwargs["stream"] is True
+
     async def test_should_append_calculator_tool_when_add_calculator_toggle_is_true(
         self, component_class, default_kwargs
     ):

src/lfx/src/lfx/components/models_and_agents/agent.py

@@ -12,6 +12,7 @@ if TYPE_CHECKING:
 from lfx.base.agents.agent import LCToolsAgentComponent
 from lfx.base.agents.default_system_prompt import DEFAULT_SYSTEM_PROMPT_TEMPLATE
 from lfx.base.agents.events import ExceptionWithMessageError
+from lfx.base.constants import STREAM_INFO_TEXT
 from lfx.base.models.unified_models import (
     get_language_model_options,
     get_llm,
@@ -211,6 +212,13 @@ class AgentComponent(ToolCallingAgentComponent):
         *LCToolsAgentComponent.get_base_inputs(),
         # removed memory inputs from agent component
         # *memory_inputs,
+        BoolInput(
+            name="stream",
+            display_name="Stream",
+            info=STREAM_INFO_TEXT,
+            value=False,
+            advanced=True,
+        ),
         BoolInput(
             name="add_current_date_tool",
             display_name="Current Date",
@@ -283,6 +291,7 @@ class AgentComponent(ToolCallingAgentComponent):
             model=self.model,
             user_id=self.user_id,
             api_key=getattr(self, "api_key", None),
+            stream=bool(getattr(self, "stream", False)),
             max_tokens=self._get_max_tokens_value(),
             watsonx_url=getattr(self, "base_url_ibm_watsonx", None),
             watsonx_project_id=getattr(self, "project_id", None),