diff --git a/Makefile b/Makefile
index 55fc994..742197f 100644
--- a/Makefile
+++ b/Makefile
@@ -22,7 +22,7 @@ install:
# --------------------------------------------------------------------------
install-frontend:
@echo "Installing frontend dependencies..."
- cd apps/frontend && npm install
+ cd apps/dashboard && npm install
@echo "Done: frontend dependencies installed."
# --------------------------------------------------------------------------
diff --git a/README.md b/README.md
index d4ae549..19092b2 100644
--- a/README.md
+++ b/README.md
@@ -7,7 +7,7 @@
[](https://gssoc.girlscript.tech)
[](https://python.org)
[](https://fastapi.tiangolo.com)
-[](https://nextjs.org)
+[](https://react.dev)
[](LICENSE)
[](CONTRIBUTING.md)
@@ -78,7 +78,7 @@ The architecture covers:
- Detection → Tracking → Temporal Memory → Reasoning pipeline
- Event-triggered VLM execution
- Redis memory design
-- FastAPI + Next.js integration
+- FastAPI + React integration
- Full component and data-flow reference
### Service Breakdown
@@ -91,7 +91,7 @@ The architecture covers:
| **VLM Layer** | LLaVA-Next / Qwen-VL | Generate natural language frame descriptions on event trigger |
| **LLM Reasoning** | Mixtral / GPT-4o / Gemini | Classify intent from caption sequence; output label + reason |
| **Backend API** | FastAPI + Celery | Async REST API; task queue for slow VLM/LLM calls |
-| **Frontend** | Next.js 14 + TypeScript | Live video, bounding box overlay, alert panel, timeline |
+| **Frontend** | React 19 + TypeScript | Live video, bounding box overlay, alert panel, timeline |
---
@@ -106,7 +106,7 @@ The architecture covers:
| LLM Reasoning | Mixtral-8x7B / GPT-4o | Configurable per cost/quality requirements |
| Backend | FastAPI + Uvicorn | Async, auto-docs, Pydantic, fastest Python API |
| Task Queue | Celery + Redis | Decouples slow VLM/LLM from real-time pipeline |
-| Frontend | Next.js 14 + TypeScript | App Router, SSE, type safety, Tailwind |
+| Frontend | React 19 + Vite | Vite dev server, SSE, type safety, Tailwind |
| Containers | Docker + docker-compose | One-command setup for all contributors |
| CI/CD | GitHub Actions | Free for open source; native GitHub integration |
| Optimization | ONNX Runtime (INT8) | 2–4× speed-up without retraining |
@@ -321,7 +321,7 @@ Request: { "alert_id": "alert_001", "correct": false, "note": "Normal employee"
| Week 4 | VLM | LLaVA-Next producing frame captions on event trigger |
| Week 5 | LLM Reasoning | Caption sequence → Suspicious/Normal + explanation |
| Week 6 | API | FastAPI live; all endpoints tested; Docker working |
-| Week 7 | Frontend | Next.js dashboard with live video, alerts, timeline |
+| Week 7 | Frontend | React dashboard with live video, alerts, timeline |
| Week 8 | Launch | Optimized, documented, CI live, 20+ GSSoC issues |
**Post-GSSoC (v2.0+):**
diff --git a/docs/ARCHITECTURE.md b/docs/ARCHITECTURE.md
index 9202151..6b54cd7 100644
--- a/docs/ARCHITECTURE.md
+++ b/docs/ARCHITECTURE.md
@@ -1,6 +1,6 @@
# Eagle Architecture Reference
-Eagle is an event-driven surveillance reasoning pipeline that converts raw video frames into natural-language risk assessments. Frames enter the detection layer (`services/detection/detector.py`), tracked entities are persisted across time (`services/tracking/tracker.py`), recent events are stored in Redis (`services/memory/memory.py`), and only meaningful behavioral changes trigger multimodal reasoning (`services/reasoning/vlm.py` + `services/reasoning/llm.py`). The final output is a structured alert served through the FastAPI backend (`apps/backend/main.py`) and visualized in the Next.js dashboard (`apps/frontend/`).
+Eagle is an event-driven surveillance reasoning pipeline that converts raw video frames into natural-language risk assessments. Frames enter the detection layer (`services/detection/detector.py`), tracked entities are persisted across time (`services/tracking/tracker.py`), recent events are stored in Redis (`services/memory/memory.py`), and only meaningful behavioral changes trigger multimodal reasoning (`services/reasoning/vlm.py` + `services/reasoning/llm.py`). The final output is a structured alert served through the FastAPI backend (`apps/backend/main.py`) and visualized in the React dashboard (`apps/dashboard/`).
---
@@ -14,7 +14,7 @@ Eagle is an event-driven surveillance reasoning pipeline that converts raw video
| VLM Captioning | LLaVA-Next / Qwen-VL | Triggered frame sequence | Natural language captions |
| LLM Reasoning | Mixtral / GPT-4o / Gemini | Caption sequence + policies | `Alert(label, confidence, reason)` |
| Backend API | FastAPI + Celery | REST requests | JSON API responses |
-| Frontend | Next.js 14 | SSE / REST payloads | Live dashboard + alert timeline |
+| Frontend | React 19 + Vite | SSE / REST payloads | Live dashboard + alert timeline |
---
@@ -38,4 +38,4 @@ F --> G[LLM Reasoning
services/reasoning/llm.py]
G --> H[FastAPI Backend
apps/backend/main.py]
-H --> I[Next.js Dashboard
apps/frontend]
\ No newline at end of file
+H --> I[React Dashboard
apps/dashboard]