Desktop mvp (#22)

* chore: logging + transcription improvements

* chore: add ax context call on rec start

* chore: amical assets

* chore: qol setup changes

* chore: add sidebar

* chore: transcriptions tab

* chore: transcriptions ui

* chore: frame improvements

* chore: ui rework

* chore logger fixes

* chore: whisper model download func

* chore: update model downloading

* chore: transcription updates

* chore: improved logging

* chore: log whisper metrics + raw pcm proc

* chore: ste up libsql

* chore: layout fixes

* chore: clean up ipcs

* chore: integrate trpc

* chore: formatting fixes

* chore: fix pnpm lock file

Author: haritabh-z01

.gitignore

@@ -36,6 +36,8 @@ yarn-error.log*
 # Misc
 .DS_Store
 *.pem
+CLAUDE.md
+.serena
 
 # Temp files
-/tmp
+/tmp

AI_INFERENCE_SETUP.md

@@ -0,0 +1,69 @@
+# AI Inference Configuration
+
+Amical supports two AI inference options for transcription:
+
+1. **OpenAI Whisper** (default)
+2. **Amical Cloud** (hosted endpoint)
+
+## Configuration
+
+### Using OpenAI (Default)
+
+By default, Amical uses OpenAI's Whisper API. You need to provide an API key either:
+
+- Through the UI: Open settings and go to "Configure API Key" tab
+- Via environment variable: `OPENAI_API_KEY=your_key_here`
+
+### Using Amical Cloud
+
+To use the Amical Cloud hosted endpoint instead:
+
+1. Set the environment variable: `USE_CLOUD_INFERENCE=true`
+2. No API key is required for cloud inference
+
+## Examples
+
+### Running with OpenAI (default)
+```bash
+# Set API key via environment
+OPENAI_API_KEY=sk-your-openai-key npm run dev
+
+# Or just run normally and set API key through UI
+npm run dev
+```
+
+### Running with Amical Cloud
+```bash
+USE_CLOUD_INFERENCE=true npm run dev
+```
+
+### Running with both environment variables
+```bash
+USE_CLOUD_INFERENCE=true OPENAI_API_KEY=sk-backup-key npm run dev
+```
+
+## Cloud Endpoint Details
+
+When using Amical Cloud (`USE_CLOUD_INFERENCE=true`), the app makes POST requests to:
+
+- **URL**: `https://dictation.amical.ai/transcribe`
+- **Headers**:
+  - `Content-Type: application/json`
+  - `x-admin-api-key: asdsa`
+  - `x-org-user-id: <random_generated_id>`
+- **Model**: Groq Whisper Large v3
+- **Format**: Returns unformatted transcription
+
+## Switching Between Modes
+
+To switch between inference modes, simply restart the application with the appropriate environment variable:
+
+```bash
+# Switch to cloud
+USE_CLOUD_INFERENCE=true npm run dev
+
+# Switch back to OpenAI
+npm run dev
+```
+
+The application will log which inference mode is being used on startup. 

apps/electron/.eslintrc.json

@@ -94,3 +94,6 @@ out/
 # Swift Build
 .build/
 bin/
+
+# VSCode
+.vscode/