feat: add LiteRT-LM setup script and update README

- Add setup.sh to download lit CLI binary and .litertlm model
- Support macOS arm64 and x86_64 architectures
- Auto-generate .env with LIT_BINARY_PATH and LIT_MODEL_PATH
- Add .gitignore for bin/, models/, .env
- Update README with Quick Setup section

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: groupthinking

mcp-servers/litert-mcp/README.md

@@ -6,6 +6,27 @@ It allows you to run inference on local models (like Gemma, Phi, Qwen) directly
 
 **Note:** This server currently wraps the `lit` CLI. Multimodal inputs (image/audio) are enabled in the interface but require the C++ API or Python bindings. The CLI wrapper currently supports **Text-only inference** until CLI flags for multimodal are verified.
 
+## Quick Setup
+
+Run the setup script to download the `lit` binary and a default model:
+
+```bash
+cd mcp-servers/litert-mcp
+./setup.sh
+```
+
+This downloads:
+- `lit` CLI binary from [google-ai-edge/LiteRT-LM](https://github.com/google-ai-edge/LiteRT-LM/releases) into `bin/`
+- A Gemma 3n model from [HuggingFace litert-community](https://huggingface.co/litert-community) into `models/`
+- Writes a `.env` file with `LIT_BINARY_PATH` and `LIT_MODEL_PATH`
+
+Then start the server:
+
+```bash
+export $(cat .env | xargs)
+python3 server.py
+```
+
 ## Prerequisites
 
 1.  **LiteRT-LM**: You must have LiteRT-LM installed or built.

mcp-servers/litert-mcp/setup.sh

@@ -0,0 +1,59 @@
+#!/usr/bin/env bash
+# Setup script for LiteRT-LM MCP server
+# Downloads the `lit` CLI binary and a small .litertlm model
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
+BIN_DIR="$SCRIPT_DIR/bin"
+MODEL_DIR="$SCRIPT_DIR/models"
+
+mkdir -p "$BIN_DIR" "$MODEL_DIR"
+
+# --- 1. Download `lit` CLI binary ---
+LIT_VERSION="${LIT_VERSION:-v0.8.1}"
+ARCH="$(uname -m)"
+case "$ARCH" in
+  arm64|aarch64) ASSET="lit-macos-arm64" ;;
+  x86_64)        ASSET="lit-macos-x86_64" ;;
+  *)             echo "Unsupported arch: $ARCH"; exit 1 ;;
+esac
+
+LIT_URL="https://github.com/google-ai-edge/LiteRT-LM/releases/download/${LIT_VERSION}/${ASSET}"
+LIT_BIN="$BIN_DIR/lit"
+
+if [ ! -x "$LIT_BIN" ]; then
+  echo "Downloading lit binary (${LIT_VERSION}, ${ASSET})..."
+  curl -fSL "$LIT_URL" -o "$LIT_BIN"
+  chmod +x "$LIT_BIN"
+  echo "✓ Downloaded to $LIT_BIN"
+else
+  echo "✓ lit binary already exists at $LIT_BIN"
+fi
+
+# --- 2. Download a small .litertlm model (Gemma 3n) ---
+MODEL_NAME="${MODEL_NAME:-gemma3n-E2B-it-int4.litertlm}"
+MODEL_URL="https://huggingface.co/litert-community/${MODEL_NAME%.litertlm}/resolve/main/${MODEL_NAME}"
+MODEL_PATH="$MODEL_DIR/$MODEL_NAME"
+
+if [ ! -f "$MODEL_PATH" ]; then
+  echo "Downloading model ${MODEL_NAME} from HuggingFace..."
+  echo "(This may take a while depending on model size)"
+  curl -fSL "$MODEL_URL" -o "$MODEL_PATH"
+  echo "✓ Downloaded to $MODEL_PATH"
+else
+  echo "✓ Model already exists at $MODEL_PATH"
+fi
+
+# --- 3. Write env file ---
+ENV_FILE="$SCRIPT_DIR/.env"
+cat > "$ENV_FILE" <<EOF
+LIT_BINARY_PATH=$LIT_BIN
+LIT_MODEL_PATH=$MODEL_PATH
+EOF
+echo "✓ Wrote $ENV_FILE"
+
+echo ""
+echo "Setup complete. Start the MCP server with:"
+echo "  cd $SCRIPT_DIR"
+echo "  export \$(cat .env | xargs)"
+echo "  python3 server.py"