OpenAgentHQ · himanshu231204 · Jul 11, 2026 · Jul 11, 2026 · Jul 11, 2026
diff --git a/.ai/00_PROJECT.md b/.ai/00_PROJECT.md
@@ -622,22 +622,6 @@ Run evaluation with multiple thresholds and JSON output.
 
 ---
 
-```bash
-oaeval ui
-```
-
-Launch interactive TUI dashboard (Textual).
-
----
-
-```bash
-oaeval ui --config config.yaml
-```
-
-Launch TUI with specific configuration.
-
----
-
 # Configuration
 
 The framework should use YAML configuration.
@@ -745,7 +729,6 @@ Avoid tightly coupling metrics, providers, and report generators.
 Version 1.0 (Stable Release):
 
 * Generic LLM-as-Judge for custom criteria
-* **Hybrid CLI UI** (Rich banner + Textual TUI dashboard)
 
 Version 2.0:
 
@@ -1111,52 +1094,3 @@ result = evaluator.evaluate(...)
 ```
 
 without maintaining two separate codebases. This architecture is scalable and aligns well with the long-term OpenAgentHQ ecosystem.
-
----
-
-# Hybrid CLI UI (v1.0 Planned)
-
-## Design Principle
-
-- **Standard CLI:** All existing commands (`oaeval evaluate`, `oaeval audit`, etc.) use Rich for beautiful output
-- **Interactive TUI:** `oaeval ui` launches Textual dashboard for power users
-- **Zero Breaking Changes:** Existing workflows continue to work unchanged
-
-## Architecture
-
-```
-┌─────────────────────────────────────────────────────────────┐
-│  oaeval evaluate  →  Beautiful Rich output (no TUI)         │
-│  oaeval audit     →  Beautiful Rich output (no TUI)         │
-│  oaeval ui        →  Full Textual interactive dashboard     │
-└─────────────────────────────────────────────────────────────┘
-```
-
-## Tech Stack Additions
-
-| Component | Technology | Reason |
-|-----------|------------|--------|
-| ASCII Art Banner | pyfiglet | Professional CLI branding |
-| Interactive TUI | Textual | Full keyboard-driven dashboard |
-
-## Dependencies
-
-```toml
-[project.optional-dependencies]
-ui = ["textual>=0.40", "pyfiglet>=1.0"]
-```
-
-## UI Module Structure
-
-```
-openagent_eval/
-├── cli/
-│   ├── banner.py        # ASCII art banner with Rich (NEW)
-│   └── commands/
-├── ui/                  # Textual TUI application (NEW)
-│   ├── __init__.py
-│   ├── app.py           # Main Textual App class
-│   ├── screens.py       # Dashboard screens (main, audit, evaluate, diagnose)
-│   ├── widgets.py       # Custom widgets (banner, results table, progress bars)
-│   └── styles.tcss      # Textual CSS for layout
-```
diff --git a/.ai/01_ARCHITECTURE.md b/.ai/01_ARCHITECTURE.md
@@ -72,19 +72,11 @@ openagent_eval/
 │   │   ├── doctor.py      # oaeval doctor
 │   │   ├── audit.py       # oaeval audit (NEW - corpus health)
 │   │   ├── diagnose.py    # oaeval diagnose (NEW - blame attribution)
-│   │   ├── synth.py       # oaeval synth (NEW - synthetic data)
-│   │   └── ui.py          # oaeval ui (NEW - Textual TUI)
+│   │   └── synth.py       # oaeval synth (NEW - synthetic data)
 │   └── utils/             # CLI utilities
 │       ├── __init__.py
 │       └── display.py     # Rich display helpers
 │
-├── ui/                     # Textual TUI application (NEW)
-│   ├── __init__.py
-│   ├── app.py             # Main Textual App class
-│   ├── screens.py         # Dashboard screens (main, audit, evaluate, diagnose)
-│   ├── widgets.py         # Custom widgets (banner, results table, progress bars)
-│   └── styles.tcss        # Textual CSS for layout
-│
 ├── config/                 # Configuration management
 │   ├── __init__.py
 │   ├── models.py          # Pydantic configuration models
@@ -266,47 +258,7 @@ def run(config_path: str):
     display_results(result)
 ```
 
-### 2. UI Layer (`ui/`) — NEW
-
-**Responsibility:** Interactive Textual TUI dashboard for power users.
-
-**Key insight:** Standard CLI commands use Rich for beautiful output. `oaeval ui` launches a full Textual dashboard for interactive exploration.
-
-**Components:**
-
-| File | Responsibility |
-|------|----------------|
-| `app.py` | Main Textual App class |
-| `screens.py` | Dashboard screens (main, audit, evaluate, diagnose) |
-| `widgets.py` | Custom widgets (banner, results table, progress bars) |
-| `styles.tcss` | Textual CSS for layout |
-
-**Example:**
-```python
-from textual.app import App
-
-class OAEvalDashboard(App):
-    """Interactive evaluation dashboard."""
-
-    def compose(self):
-        yield Header()
-        yield DashboardScreen()
-        yield Footer()
-
-    BINDINGS = [
-        ("1", "run_evaluation", "Run Eval"),
-        ("2", "audit_corpus", "Audit Corpus"),
-        ("q", "quit", "Quit"),
-    ]
-```
-
-**Rules:**
-- TUI is opt-in via `oaeval ui` command
-- Standard CLI commands remain unchanged
-- No business logic in UI layer
-- Display-only (delegates to core modules)
-
-### 3. Core Orchestration (`core/`)
+### 2. Core Orchestration (`core/`)
 
 **Responsibility:** Orchestrate the evaluation pipeline.
 
@@ -553,13 +505,11 @@ cli/ → core/ → datasets/
              → corpus/      (NEW)
              → diagnosis/   (NEW)
              → synthesis/   (NEW)
-ui/ → core/  (same dependencies as cli/)
 ```
 
 **Rules:**
 1. `cli/` depends on everything
-2. `ui/` depends on `core/` (same as `cli/`)
-3. `core/` depends on `datasets/`, `metrics/`, `providers/`, `reports/`, `corpus/`, `diagnosis/`, `synthesis/`
+2. `core/` depends on `datasets/`, `metrics/`, `providers/`, `reports/`, `corpus/`, `diagnosis/`, `synthesis/`
 3. `metrics/`, `providers/`, `reports/` depend only on `utils/` and `types/`
 4. `corpus/` depends on `providers/` (for LLM-as-Judge) and `utils/`
 5. `diagnosis/` depends on `metrics/` (for metric results) and `utils/`

diff --git a/.ai/02_AGENT.md b/.ai/02_AGENT.md
@@ -525,45 +525,6 @@ class MetricResult:
 
 ---
 
-## Hybrid CLI Architecture (Implemented - Phase 14)
-
-### Design Principle
-
-- **Standard CLI:** All existing commands (`oaeval evaluate`, `oaeval audit`, etc.) use Rich for beautiful output
-- **Interactive TUI:** `oaeval ui` launches Textual dashboard for power users
-- **Zero Breaking Changes:** Existing workflows continue to work unchanged
-
-### Tech Stack Additions
-
-| Component | Technology | Reason |
-|-----------|------------|--------|
-| ASCII Art Banner | pyfiglet | Professional CLI branding |
-| Interactive TUI | Textual | Full keyboard-driven dashboard |
-
-### Dependencies
-
-```toml
-[project.optional-dependencies]
-ui = ["textual>=0.40", "pyfiglet>=1.0"]
-```
-
-### UI Module Structure
-
-```
-openagent_eval/
-├── cli/
-│   ├── banner.py        # ASCII art banner with Rich
-│   └── commands/
-├── ui/                  # Textual TUI application
-│   ├── __init__.py
-│   ├── app.py           # Main Textual App class
-│   ├── screens.py       # Dashboard screens (main, audit, evaluate, diagnose)
-│   ├── widgets.py       # Custom widgets (banner, results table, progress bars)
-│   └── styles.tcss      # Textual CSS for layout
-```
-
----
-
 ## Development Phases
 
 ### Phase 1: Foundation ✅
@@ -657,15 +618,6 @@ openagent_eval/
 - [x] Write documentation for CI/CD integration
 - [x] Add GitHub Actions workflow example
 
-### Phase 14: Hybrid CLI UI ✅
-- [x] Add `pyfiglet` and `textual` to optional dependencies
-- [x] Create `openagent_eval/cli/banner.py` — ASCII art banner with Rich
-- [x] Create `openagent_eval/ui/` module structure
-- [x] Implement Textual dashboard app
-- [x] Create dashboard screens (main, audit, evaluate, diagnose)
-- [x] Add custom widgets (banner, results table, progress bars)
-- [x] Wire up `oaeval ui` command
-
 ---
 
 ## AI Library Dependencies

diff --git a/.ai/03_CONTEXT.md b/.ai/03_CONTEXT.md
@@ -9,11 +9,11 @@
 
 | Field | Value |
 |-------|-------|
-| **Phase** | Phase 14 Complete - Hybrid CLI UI |
-| **Status** | v1.0 complete; production-grade features in progress |
-| **Last Updated** | 2026-07-11 |
+| **Phase** | Phase 14 Complete - TUI Removed |
+| **Status** | v1.0 complete; CLI-only interface |
+| **Last Updated** | 2026-07-12 |
 | **Next Action** | Another feature or v1.0 release |
-| **Current Branch** | feature/hybrid-cli-ui |
+| **Current Branch** | docs/remove-tui-references |
 | **Remote** | https://github.com/OpenAgentHQ/openagent-eval.git |
 
 ---
@@ -224,6 +224,7 @@ SDK (openagent_eval - Core Evaluation API)
 - [x] Wire up `oaeval ui` command
 - [x] Add keyboard shortcuts and navigation
 - [x] Test and polish (15 tests passing)
+- [x] **REMOVED:** TUI removed from codebase, CLI-only interface retained
 
 ---
 
@@ -292,7 +293,7 @@ chore/{description}        # Maintenance tasks
 - Phase 8 is pending - Documentation
 - **Phase 9-13 are NEW** — Production-grade RAG evaluation features
 - **Phase 13 is COMPLETE** — CI/CD Integration (pytest plugin, threshold gating, `oaeval test` command)
-- **Phase 14 is COMPLETE** — Hybrid CLI UI (Rich banner, Textual TUI dashboard, 15 tests)
+- **Phase 14 is COMPLETE** — TUI removed, CLI-only interface retained
 - **Phase 12 is COMPLETE** — Synthetic Test Data generator implemented (56 tests)
 - CORRECTION: earlier "517+ passing / all phases complete" status was inaccurate —
   the core pipeline did not actually evaluate. That gap is closed.
@@ -309,36 +310,17 @@ chore/{description}        # Maintenance tasks
 3. **NLI-based Scoring** — Word overlap is insufficient for production use
 4. **Synthetic Test Data** — Bootstrap evaluation from your corpus
 5. **CLI-first, Local-first** — No cloud services, no dashboards, no authentication
-6. **Hybrid CLI UI** — Beautiful Rich output + optional Textual TUI dashboard (IMPLEMENTED)
-
----
-
-## CLI UI Research Findings
-
-**Key technologies discovered:**
-
-| Technology | Purpose | Integration |
-|------------|---------|-------------|
-| pyfiglet | ASCII art banners | CLI banner generation |
-| Rich Layout | Rows/columns, Panels, Tables | Standard CLI output |
-| Textual | Full TUI framework | `oaeval ui` command |
-| rich.live.Live | Real-time dashboards | Live updating displays |
-
-**Architecture decision:**
-- Hybrid approach: Rich for all commands, Textual for `oaeval ui`
-- Zero breaking changes to existing CLI
-- Opt-in interactivity for power users
 
 ---
 
 ## Change Log
 
 | Date | Change |
 |------|--------|
+| 2026-07-12 | **TUI REMOVED** — Textual TUI dashboard removed, CLI-only interface retained |
 | 2026-07-11 | **Phase 14 COMPLETE** — Hybrid CLI UI implemented (Rich banner, Textual TUI dashboard, 15 tests) |
 | 2026-07-11 | **Phase 13 COMPLETE** — CI/CD Integration implemented (pytest plugin, threshold gating, `oaeval test` command) |
 | 2026-07-11 | **Phase 12 COMPLETE** — Synthetic Test Data generator implemented (56 tests) |
-| 2026-07-11 | Added Phase 14: Hybrid CLI UI (Rich banner + Textual TUI dashboard) |
 | 2026-07-11 | Added CLI UI research findings to context |
 | 2026-07-11 | Added Phase 9-13 (production-grade RAG eval): Corpus Auditor, LLM-as-Judge, Diagnosis, Synthetic Data, CI/CD |
 | 2026-07-11 | Updated 00_PROJECT.md with production-grade vision and feature matrix |

diff --git a/.ai/04_DECISIONS.md b/.ai/04_DECISIONS.md
@@ -590,35 +590,6 @@
 
 ---
 
-### D022: Hybrid CLI UI Architecture
-
-**Date:** 2026-07-11
-**Status:** Accepted
-**Context:** Need beautiful CLI output without breaking existing workflows. Users want both quick CLI evaluations and optional interactive dashboards.
-
-**Decision:** Hybrid approach — Rich banner for all commands, Textual TUI via `oaeval ui`.
-
-**Rationale:**
-- Zero breaking changes to existing CLI workflows
-- Opt-in interactivity for power users
-- Beautiful ASCII art banner for professional branding
-- Full Textual dashboard for interactive exploration
-- Best of both worlds: simple CLI + powerful TUI
-
-**Consequences:**
-- New `ui/` module for Textual application
-- New `cli/banner.py` for ASCII art banner
-- New optional dependencies: `textual`, `pyfiglet`
-- Standard CLI commands remain unchanged
-- `oaeval ui` launches interactive dashboard
-
-**Alternatives Considered:**
-- Replace entire CLI with Textual: Rejected, breaking change, higher complexity
-- Rich formatting only: Rejected, no interactive mode
-- Textual for all commands: Rejected, overkill for simple evaluations
-
----
-
 ## Pending Decisions
 
 None at this time.