add config app

README.md

@@ -153,6 +153,16 @@ eva \
     --metrics task_completion,faithfulness,conciseness
 ```
 
+### Configuring EVA
+
+EVA includes a Streamlit config editor for building your `.env` file interactively:
+
+```bash
+streamlit run apps/config_editor.py
+```
+
+The editor covers all variables grouped by tab (API keys, voice pipeline, model deployments, runtime settings, perturbations, etc.), with proper widgets for each type. See [`apps/README.md`](apps/README.md) for details.
+
 ### Exploring Results
 
 EVA includes a Streamlit analysis app for visualizing and comparing results:

apps/README.md

@@ -1,6 +1,20 @@
 # EVA Apps
 
-Streamlit applications for exploring EVA results.
+Streamlit applications for exploring and configuring EVA.
+
+## Config Editor
+
+Interactive UI for building and editing `.env` configuration files without hand-editing JSON or looking up variable names.
+
+### Usage
+
+```bash
+streamlit run apps/config_editor.py
+```
+
+The app reads `.env.example` for the full variable set and loads existing values from `.env` if present. Each variable's widget type, enum options, ranges, and tooltips are declared directly in `.env.example` using annotation prefixes (`#i`, `#d`, `#e`, `#r`, `#x`, `#v`). Use the **Preview** button to inspect the generated file before saving, or **Download** to export it without writing to disk.
+
+---
 
 ## Analysis App
 

apps/analysis.py

@@ -1585,7 +1585,7 @@ def _pivot_metric_label(m: str) -> str:
                 height=max(350, 80 + 40 * len(y_labels)),
                 margin={"l": 20, "r": 20, "t": 50, "b": 120},
             )
-            st.plotly_chart(heatmap_fig, use_container_width=True)
+            st.plotly_chart(heatmap_fig, width="stretch")
 
 
 def render_run_overview(run_dir: Path):