# Train a LoRA → mention trigger word → compose multiple LoRAs

1. **Train:** start a LoRA training job. Two entry points — the **dashboard** LoRA training surface, or the in-canvas **AI Tools → LoRA Trainer**. You need at least 10 images; pick **style** or **object** training, and Framo generates (or you set) a **trigger word**. See [Training & using LoRAs](/docs/product/lora).
2. **Pick the LoRA in chat:** with the category on **Image** (or **Enhance**), open the **LoRA** pill → **Pick a LoRA** to choose one from your library. You get the trigger word plus a **scale** knob (scroll or click for a slider; ~1.00 baseline).
3. **Invoke the trigger word:** type the trigger word in your prompt to activate the adapter. You can also type **`@`** in the prompt to mention a LoRA inline by its trigger word.
4. **Compose multiple:** add more than one LoRA the same way to stack a character/style combination in a single generation. See [Chat bar: generating images](/docs/product/chat-image-gen).