diff --git a/tests/app/routers/test_qa.py b/tests/app/routers/test_qa.py
new file mode 100644
index 0000000..40f11e9
--- /dev/null
+++ b/tests/app/routers/test_qa.py
@@ -0,0 +1,167 @@
+from unittest.mock import AsyncMock, Mock
+
+import pytest
+
+from neuroagent.app.config import Settings
+from neuroagent.app.dependencies import get_agents_routine, get_settings
+from neuroagent.app.main import app
+from neuroagent.app.routers import qa
+from neuroagent.new_types import Agent, Response
+
+
+@pytest.mark.httpx_mock(can_send_already_matched_responses=True)
+def test_run_agent(app_client, patch_required_env, httpx_mock, db_connection):
+    agent_output = Response(
+        messages=[
+            {"role": "user", "content": "Hello"},
+            {"content": "Hello! How can I assist you today?"},
+        ],
+        agent=Agent(
+            name="Agent",
+            model="gpt-4o-mini",
+            instructions="You are a helpfull assistant",
+        ),
+    )
+    agent_routine = AsyncMock()
+    agent_routine.arun.return_value = agent_output
+
+    test_settings = Settings(
+        db={"prefix": db_connection},
+    )
+    app.dependency_overrides[get_settings] = lambda: test_settings
+    app.dependency_overrides[get_agents_routine] = lambda: agent_routine
+    httpx_mock.add_response(
+        url=f"{test_settings.virtual_lab.get_project_url}/32c83739-f39c-49d1-833f-58c981ebd2a2/projects/123251a1-be18-4146-87b5-5ca2f8bfaf48"
+        # Not great to hardcode the Vlab and Project
+    )
+    with app_client as app_client:
+        response = app_client.post("/qa/run", json={"query": "This is my query"})
+
+        assert response.status_code == 200
+        assert (
+            response.json()["message"]
+            == agent_output.model_dump()["messages"][-1]["content"]
+        )
+
+        # Missing query
+        response = app_client.post("/qa/run", json={})
+        assert response.status_code == 422
+
+
+@pytest.mark.httpx_mock(can_send_already_matched_responses=True)
+def test_run_chat_agent(app_client, httpx_mock, patch_required_env, db_connection):
+    agent_output = Response(
+        messages=[
+            {"role": "user", "content": "Hello"},
+            {"content": "Hello! How can I assist you today?", "role": "assistant"},
+        ],
+        agent=Agent(
+            name="Agent",
+            model="gpt-4o-mini",
+            instructions="You are a helpfull assistant",
+        ),
+    )
+    agent_routine = AsyncMock()
+    agent_routine.arun.return_value = agent_output
+
+    test_settings = Settings(
+        db={"prefix": db_connection},
+    )
+    app.dependency_overrides[get_settings] = lambda: test_settings
+    app.dependency_overrides[get_agents_routine] = lambda: agent_routine
+    httpx_mock.add_response(
+        url=f"{test_settings.virtual_lab.get_project_url}/test_vlab/projects/test_project"
+    )
+    with app_client as app_client:
+        create_output = app_client.post(
+            "/threads/?virtual_lab_id=test_vlab&project_id=test_project"
+        ).json()
+        response = app_client.post(
+            f"/qa/chat/{create_output['thread_id']}",
+            json={"query": "This is my query"},
+            headers={"x-virtual-lab-id": "test_vlab", "x-project-id": "test_project"},
+        )
+        assert response.status_code == 200
+        assert (
+            response.json()["message"]
+            == agent_output.model_dump()["messages"][-1]["content"]
+        )
+
+        # Missing thread_id query
+        response = app_client.post(
+            "/qa/chat",
+            json={"query": "This is my query"},
+            headers={"x-virtual-lab-id": "test_vlab", "x-project-id": "test_project"},
+        )
+        assert response.status_code == 404
+
+
+async def streamed_response():
+    response = [
+        "Calling ",
+        "tool ",
+        ": ",
+        "resolve_entities_tool ",
+        "with ",
+        "arguments ",
+        ": ",
+        "{",
+        "brain_region",
+        ": ",
+        "thalamus",
+        "}",
+        "\n ",
+        "This",
+        " is",
+        " an",
+        " amazingly",
+        " well",
+        " streamed",
+        " response",
+        ".",
+        " I",
+        " can",
+        "'t",
+        " believe",
+        " how",
+        " good",
+        " it",
+        " is",
+        "!",
+    ]
+    for word in response:
+        yield word
+
+
+@pytest.mark.httpx_mock(can_send_already_matched_responses=True)
+def test_chat_streamed(app_client, httpx_mock, patch_required_env, db_connection):
+    """Test the generative QA endpoint with a fake LLM."""
+    qa.stream_agent_response = Mock()
+    qa.stream_agent_response.return_value = streamed_response()
+
+    test_settings = Settings(
+        db={"prefix": db_connection},
+    )
+    app.dependency_overrides[get_settings] = lambda: test_settings
+    agent_routine = Mock()
+    app.dependency_overrides[get_agents_routine] = lambda: agent_routine
+
+    expected_tokens = (
+        b"Calling tool : resolve_entities_tool with arguments : {brain_region:"
+        b" thalamus}\n This is an amazingly well streamed response. I can't believe how"
+        b" good it is!"
+    )
+    httpx_mock.add_response(
+        url=f"{test_settings.virtual_lab.get_project_url}/test_vlab/projects/test_project"
+    )
+    with app_client as app_client:
+        create_output = app_client.post(
+            "/threads/?virtual_lab_id=test_vlab&project_id=test_project"
+        ).json()
+        response = app_client.post(
+            f"/qa/chat_streamed/{create_output['thread_id']}",
+            json={"query": "This is my query"},
+            headers={"x-virtual-lab-id": "test_vlab", "x-project-id": "test_project"},
+        )
+    assert response.status_code == 200
+    assert response.content == expected_tokens