# LocalAI v2.22.1 — Release & Upgrade Intelligence
- **Product:** LocalAI (https://localai.io)
- **Version:** v2.22.1
- **Released:** Oct 21, 2024
- **Impact Score:** 5/10
- **Breaking Changes:** Not verified
- **Canonical Analysis:** https://changelogs.us/localai/updates/v2.22.1/

## Executive Summary (TL;DR)
LocalAI v2.22.1. Source highlight: fix(vllm): images and videos are base64 by default

## New Features
- feat(templates): extract text from multimodal requests
- feat(templates): add sprig to multimodal templates
- models(gallery): add llama-3_8b_unaligned_beta
- models(gallery): add llama3.1-flammades-70b
- models(gallery): add llama3.1-gutenberg-doppel-70b
- models(gallery): add llama-3.1-8b-arliai-formax-v1.0-iq-arm-imatrix

## Bug Fixes
- fix(vllm): images and videos are base64 by default
- fix(dependencies): pin pytorch version
- fix(dependencies): move deps that brings pytorch
- fix(vllm): do not set videos if we don't have any
- models(gallery): add mn-lulanum-12b-fix-i1
- chore(deps): Bump debugpy from 1.8.6 to 1.8.7 in /examples/langchain/langchainpy-localai-example