docs: remove outdated Metal compile lockup warning as MoE streamed inference primitives resolve the delay

Author: solderzzc

README.md

@@ -116,11 +116,6 @@ When your agent engine spins up `mlx-server` using standard subprocess managemen
 > **Critical Memory Routing (`--stream-experts`)**  
 > When managing MoE models (e.g., `Qwen3.5-122B-A10B` where active parameters are significantly smaller than the total payload), you **must artificially append** the `--stream-experts true` flag to the process arguments. If omitted, macOS will inevitably suffer a `Data Abort` memory-mapping fault when mapping > 100GB of tensors onto the unified GPU hardware.
 
-### 3. Client Reliability Overrides (First-Request Lock)
-> [!WARNING]
-> Because zero-copy heavy matrices (like 122B parameter graphs) require **3-5+ minutes of pure compile-time lockup** on their *very first request* to build the specialized Apple Metal compilation graphs, **standard short-timeouts (e.g., 60s) will fail the first request.**  
-> 
-> Ensure your system's `node-fetch` metrics checks, network controllers, or `curl` abort signals extend standard timeouts to `> 300s`.
 
 ---
 ## ⚙️ CLI Options

mlx_integration_prompt.md

@@ -0,0 +1,37 @@
+# Aegis-AI Integration Prompt: mlx-server Backend
+
+Aegis-AI can seamlessly substitute `llama-server` or cloud VLM gateways with `mlx-server` for dramatically faster zero-latency inference on macOS instances.
+
+When configuring local inference within the `~/.aegis-ai/llm-config.json` configuration layer, apply the following integration details:
+
+## 1. Gateway Emulation
+`mlx-server` exposes a fully standard OpenAI-compatible API layer. 
+- **Endpoint**: `http://127.0.0.1:5413/v1` (or whatever `port` is provided)
+- **Supported Routes**:
+  - `GET /health` (Exposes advanced JSON containing GPU metrics/VRAM allocations)
+  - `GET /v1/models`
+  - `POST /v1/chat/completions` (Supports both `stream: true` and `stream: false`, natively handles tool-calls, system prompts, and temperature checks)
+
+## 2. Process Configuration (The Sidecar)
+
+The Aegis-AI deployment engine should spin up the `mlx-server` sidecar using standard subprocess management. 
+
+### Launch Command
+```bash
+/path/to/mlx-server/.build/release/mlx-server \
+  --model {MODEL_IDENTIFIER_OR_PATH} \
+  --host 127.0.0.1 \
+  --port {PORT_NUMBER}
+```
+
+### 🧠 Critical Memory Routing Flag (`--stream-experts`)
+When managing MoE models (e.g. `Qwen3.5-122B-A10B` or any model where active parameters (`A10B`) are substantially smaller than totally memory payload), Aegis-AI **must artificially append** the `--stream-experts true` flag to the process arguments. 
+If this flag is omitted, macOS will inevitably suffer a `Data Abort` memory-mapping fault when loading > 100GB of tensors onto the unified GPU hardware. 
+
+Example:
+```bash
+.build/release/mlx-server --model mlx-community/Qwen3.5-122B-A10B-4bit --stream-experts true --port 5413
+```
+
+
+_You can safely feed this document block sequentially into the Aegis zero-assumption onboarding agent to establish permanent MLX local pipelines._