User Workflow

For Everyday Users

Step 1: Install Ego Pod
  $ pip install ego-pod

Step 2: Initialize Configuration
  $ ego init
  → Interactive guide: select base model, configure data source paths, choose editions
  → Like setting up a new phone, just follow the prompts step by step

Step 3: Start Training
  $ ego train
  → Auto data collection → corpus building → LoRA training → multi-version Ego output
  → In the time it takes to drink a coffee, your personalized AI is ready

Step 4: Use Ego
  $ ego run --edition normal
  → Load base + LoRA, start an AI model aligned with you
  → Try chatting with it, see if it "sounds like you"

Step 5: Update Ego (after data changes)
  $ ego train --update
  → Incremental data collection → retrain → overwrite old version
  → Learned something new recently? Forge again, and Ego catches up

iFay Integration

The iFay body tells Ego the data source addresses through configuration files. Ego Pod automatically completes training, and the produced LoRA adapters are loaded by Fayger.

For example: Your iFay runs on your phone, periodically syncing new data to a designated directory, then triggering Ego Pod to retrain. After training completes, the new LoRA adapter is automatically pushed to your phone, and Ego "evolves" — the entire process is invisible to you.