diff --git a/backends/stable_diffusion_torch/models/ldm/stable-diffusion-v1/model.ckpt b/backends/stable_diffusion_torch/models/ldm/stable-diffusion-v1/model.ckpt
deleted file mode 120000
index 174e5d9b..00000000
--- a/backends/stable_diffusion_torch/models/ldm/stable-diffusion-v1/model.ckpt
+++ /dev/null
@@ -1 +0,0 @@
-/Users/divamgupta/Downloads/sd-v1-4.ckpt
\ No newline at end of file
diff --git a/electron_app/package-lock.json b/electron_app/package-lock.json
index f2451c53..b8d28399 100644
--- a/electron_app/package-lock.json
+++ b/electron_app/package-lock.json
@@ -1,12 +1,12 @@
 {
   "name": "DiffusionBee",
-  "version": "0.1.0",
+  "version": "0.3.0",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "DiffusionBee",
-      "version": "0.1.0",
+      "version": "0.3.0",
       "hasInstallScript": true,
       "dependencies": {
         "@codekraft-studio/vue-record": "^0.0.3",
diff --git a/electron_app/src/bridge.js b/electron_app/src/bridge.js
index d6551f09..2e19d0ef 100644
--- a/electron_app/src/bridge.js
+++ b/electron_app/src/bridge.js
@@ -12,7 +12,9 @@ function start_bridge() {
     console.log("starting briddddd")
     const fs = require('fs')
 
-    let script_path = process.env.PY_SCRIPT || "./src/fake_backend.py"; 
+    // PY_SCRIPT_DEV can be set to "./src/fake_backend.py" when developing the electron app
+    // to simulate the model inference, without actually executing the code
+    let script_path = process.env.PY_SCRIPT_DEV || "../backends/stable_diffusion_torch/txt2img.py";
 
     if (fs.existsSync(script_path)) {
         python = require('child_process').spawn('python3', [script_path]);
@@ -129,4 +131,4 @@ function bind_window_bridge(w) {
 }
 
 
-export { start_bridge, bind_window_bridge }
\ No newline at end of file
+export { start_bridge, bind_window_bridge }
diff --git a/install-local-dev-2.sh b/install-local-dev-2.sh
new file mode 100755
index 00000000..069f8bb4
--- /dev/null
+++ b/install-local-dev-2.sh
@@ -0,0 +1,20 @@
+#!/bin/bash
+set -e
+
+cd backends/stable_diffusion_torch/
+
+echo ">> Installing pytorch"
+conda install pytorch torchvision torchaudio -c pytorch-nightly
+
+echo ">> Downloading custom python dependencies, might have to pass (i) to ignore existing packages"
+pip install -r requirements.txt
+
+cd ../..
+
+echo ">> Setting up electron app"
+cd electron_app
+npm i
+
+echo ">> Starting electron app"
+npm run electron:serve
+
diff --git a/install-local-dev.sh b/install-local-dev.sh
new file mode 100755
index 00000000..cde67299
--- /dev/null
+++ b/install-local-dev.sh
@@ -0,0 +1,34 @@
+#!/bin/bash
+
+set -e
+
+# Download stable diffusion info
+cd backends/stable_diffusion_torch
+
+echo ">> Downloading stable diffusion code"
+# Download the m1-compatible Stable Diffusion sampling infrastructure
+git clone https://github.com/magnusviri/stable-diffusion.git
+cd stable-diffusion
+git checkout apple-mps-support
+cd ..
+
+echo ">> Downloading text (CLIP) tokenizer"
+# Download the CLIP tokenizer from huggingface hub
+git clone https://huggingface.co/openai/clip-vit-base-patch32
+mkdir -p HF_weights/
+mv clip-vit-base-patch32 HF_weights/clip_tokenizer
+
+echo ">> Downloading model weights (4gb, may take awhile)"
+# Download v1.4 model weights (so you can run `python txt2img.py` locally when testing)
+wget "https://me.cmdr2.org/stable-diffusion-ui/sd-v1-4.ckpt" -O sd-v1-4.ckpt
+mkdir -p models/ldm/stable-diffusion-v1/
+ln -s sd-v1-4.ckpt models/ldm/stable-diffusion-v1/model.ckpt
+
+
+echo ">> Creating environment diffusion bee"
+# Create fresh python environment
+conda create --name diffusionbee
+
+echo ">> Now run 'conda activate diffusionbee'"
+echo ">> Followed by './install-local-dev-2.sh'" 
+