diff --git a/docs/readme-redesign-proposal.md b/docs/readme-redesign-proposal.md
new file mode 100644
index 0000000..9d6cfa8
--- /dev/null
+++ b/docs/readme-redesign-proposal.md
@@ -0,0 +1,408 @@
+
+Symphony-Coord
+
+
+
+Agents That Learn Who Should Solve What
+
+
+
+Self-Organizing Multi-Agent Coordination via Adaptive Online Routing
+
+
+
+
+ 
+
+ 
+ 
+
+
+
+ ๐ Paper
+ ยท
+ ๐ Live Demo
+ ยท
+ ๐ก Ecosystem
+
+
+---
+
+
+
+ 
+
+
+
+
+Decentralized agents that dynamically learn who should solve what.
+
+
+---
+
+# Overview
+
+Symphony-Coord is a decentralized multi-agent coordination framework where agents dynamically learn:
+
+- who should solve what
+- when to route tasks
+- how to specialize through interaction
+
+Instead of relying on fixed orchestration heuristics or centralized planners, Symphony-Coord formulates routing as an online decision-making problem under uncertainty.
+
+Routing policies continuously evolve through:
+
+- contextual online routing
+- reward-driven adaptation
+- decentralized coordination
+- emergent specialization
+
+The framework is designed for dynamic environments where:
+
+- agent capability changes over time
+- latency fluctuates
+- nodes fail or degrade
+- specialization must emerge online
+
+---
+
+# Why Symphony-Coord?
+
+Modern multi-agent systems often rely on:
+
+- centralized orchestrators
+- static expert assignment
+- fixed routing heuristics
+
+However, real-world decentralized systems are inherently dynamic.
+
+Agent capability, latency, availability, and specialization continuously evolve during execution.
+
+Symphony-Coord studies how robust coordination and specialization can emerge through online interaction instead of predefined orchestration rules.
+
+---
+
+# System Demo
+
+Explore adaptive routing and emergent specialization in decentralized multi-agent systems.
+
+
+
+ 
+
+
+
+---
+
+# Dynamic System Behavior
+
+## Adaptive Routing Evolution
+
+
+ 
+
+
+Routing decisions evolve online as task streams and reward feedback continuously change.
+
+---
+
+## Emergent Agent Specialization
+
+
+ 
+
+
+Agents gradually specialize through interaction and reward feedback instead of predefined static roles.
+
+---
+
+## Robust Failure Recovery
+
+
+ 
+
+
+The system dynamically adapts after agent degradation, routing disruption, or node failure.
+
+---
+
+# Main Results
+
+| Evaluation Setting | Improvement |
+|---|---:|
+| Routing Cost vs. Static Routing | โ 23% |
+| Recovery Speed under Agent Failure | โ 2.1ร |
+| GSM8K Accuracy vs. Routing Baseline | โ 8.4% |
+
+> Evaluated across GSM8K, BBH, robustness recovery, and heterogeneous system optimization benchmarks.
+
+---
+
+# Interactive Demo
+
+Explore adaptive routing and emergent specialization in real time.
+
+
+
+ 
+
+
+
+### Interactive Features
+
+- live routing visualization
+- evolving specialization dynamics
+- decentralized coordination simulation
+- adaptive recovery under failure
+- multi-agent execution tracing
+
+---
+
+# Core Features
+
+## ๐ง Emergent Specialization
+
+Agents dynamically specialize through online interaction and reward feedback.
+
+No predefined expert assignment is required.
+
+---
+
+## โก Adaptive Online Routing
+
+Routing decisions continuously evolve using contextual bandit optimization and online reward estimation.
+
+---
+
+## ๐ Decentralized Coordination
+
+No centralized orchestration bottleneck.
+
+Agents coordinate through distributed routing and capability-aware interaction.
+
+---
+
+## ๐ Robust Failure Recovery
+
+The framework adapts under:
+
+- unavailable agents
+- degraded performance
+- latency shifts
+- dynamic workloads
+
+---
+
+## ๐ Parallel Multi-Path Reasoning
+
+Symphony-Coord combines:
+
+- decentralized routing
+- parallel Chain-of-Thought execution
+- voting-based aggregation
+
+for robust multi-agent reasoning.
+
+---
+
+# System Architecture
+
+
+ 
+
+
+Symphony-Coord follows a three-stage coordination pipeline.
+
+---
+
+## 1. Planning
+
+๐งฉ Task decomposition and candidate plan generation.
+
+Core components:
+
+- task decomposition
+- plan proposal generation
+- uncertainty-aware plan selection
+
+---
+
+## 2. Adaptive Routing
+
+๐ Decentralized capability-aware coordination.
+
+Core components:
+
+- contextual routing
+- capability matching
+- online reward adaptation
+- emergent specialization
+
+---
+
+## 3. Voting & Aggregation
+
+๐ง Robust multi-path reasoning fusion.
+
+Core components:
+
+- parallel CoT execution
+- confidence estimation
+- voting-based aggregation
+- final answer fusion
+
+---
+
+# Quick Start
+
+## Installation
+
+```bash
+git clone https://github.com/GradientHQ/symphony-coord.git
+cd symphony-coord
+
+python -m venv venv
+source venv/bin/activate
+
+pip install --upgrade pip
+pip install -r requirements.txt
+pip install -e .
+````
+
+---
+
+## Configure API Key
+
+```bash
+export OPENROUTER_API_KEY="your-key"
+```
+
+---
+
+## Run Example
+
+```python
+from symphony import SymphonyOrchestrator
+
+orchestrator = SymphonyOrchestrator(
+ agents=["agent1", "agent2", "agent3"],
+ topL=3,
+ cot_count=3,
+)
+
+result = orchestrator.run_task(
+ task_description="Solve: What is 25 * 37?",
+ requirements=["math"],
+)
+
+print(result["final_answer"])
+```
+
+---
+
+# Reproducing Results
+
+Run the benchmark suite:
+
+```bash
+bash experiments/scripts/run_all_datasets.sh
+```
+
+Generate paper figures:
+
+```bash
+python scripts/plotting/paper_figures/plot_robustness_bars.py
+python scripts/plotting/paper_figures/plot_gap_analysis.py
+```
+
+---
+
+# Ecosystem
+
+Explore the Symphony-Coord ecosystem.
+
+### Resources
+
+* ๐ Interactive system demo
+* ๐ก Research discussions
+* ๐ Routing and specialization visualization
+* ๐ Open experiments and extensions
+
+### Links
+
+* [GradientHQ](https://github.com/GradientHQ)
+* [GitHub Discussions](https://github.com/GradientHQ/symphony-coord/discussions)
+* [Issues](https://github.com/GradientHQ/symphony-coord/issues)
+
+---
+
+# Roadmap
+
+* [ ] Interactive routing visualization
+* [ ] Dynamic specialization analysis
+* [ ] Multi-node distributed deployment
+* [ ] Real-time coordination dashboard
+* [ ] Open benchmark suite
+* [ ] Agent memory and long-horizon coordination
+
+---
+
+# Documentation
+
+Detailed setup and experiment guides are available in:
+
+```text
+docs/
+โโโ INSTALL.md
+โโโ EXPERIMENTS.md
+โโโ CONFIGS.md
+โโโ TROUBLESHOOTING.md
+โโโ OPENROUTER_CONFIG_GUIDE.md
+```
+
+---
+
+# Repository Structure
+
+```text
+symphony-coord/
+โโโ agents/ # Agent implementations
+โโโ core/ # Routing and coordination algorithms
+โโโ experiments/ # Benchmark and robustness experiments
+โโโ protocol/ # Task and beacon protocols
+โโโ scripts/ # Plotting and analysis scripts
+โโโ docs/ # Documentation
+โโโ tests/ # Test suite
+โโโ symphony.py # Main orchestrator
+```
+
+---
+
+# Citation
+
+```bibtex
+@article{guan2026symphony,
+ title={Symphony-Coord: Emergent Coordination in Decentralized Agent Systems},
+ author={Guan, Zhaoyang and Cao, Huixi and Zhong, Ming and Yang, Eric and Ai, Lynn and Ni, Yongxin and Shi, Bill},
+ journal={arXiv preprint arXiv:2602.00966},
+ year={2026}
+}
+```
+
+---
+
+# Acknowledgements
+
+We thank the open-source research community for foundational work in:
+
+* decentralized systems
+* online bandit optimization
+* multi-agent reasoning
+* Chain-of-Thought coordination
+* distributed inference systems
+
+---
+
+# License
+
+MIT License