Hey everyone! Quick question - I’m trying to send an image from Form Trigger to my self-hosted LLaVA (on Ollama/Railway) for OCR, but I can’t figure out how to get the binary data to actually reach the vision model. The form captures the file metadata fine, but when I try to extract the binary with item.binary[key].data.toString('base64') and send it to Ollama’s /api/generate endpoint, LLaVA doesn’t seem to “see” the image. Should I be using AI Agent nodes or just HTTP Request? Has anyone got Form uploads working with vision models? Thanks! 
Hi @Kaouthar_Bousbaa ,
Use HTTP Request Node
Your workflow should be like this;
Form Trigger → Code Node (prepare image) → HTTP Request (to Ollama)
Why dont you just use the Ollama node for this. Connect your auth to the Ollama instance you have and then use the analyze image with the llava model selected?