mempool/umbrel-apps
mempool/umbrel-apps
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update localai to version v4.0.0 (#5099)

Browse Source
This commit is contained in:
Alexander L. 2026-03-17 14:54:24 +01:00 committed by GitHub
parent 59b2b9a44e
commit 00c53fc8d3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 14 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
localai/docker-compose.yml
View File

@ -7,7 +7,7 @@ services:
APP_PORT: 8080
PROXY_AUTH_WHITELIST: "/v1/*"
api:
image: localai/localai:v3.12.1@sha256:558efc7610e2ce43e702eaed4fcf456102a416471929d99819ef251fdbbf8b51
image: localai/localai:v4.0.0@sha256:b841e2bef58b6277b2024726e18b01d2067c8fa44e47c687fd22e41044a4efb0
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8080/readyz"]
interval: 1m

22
localai/umbrel-app.yml
View File

@ -2,7 +2,7 @@ manifestVersion: 1
id: localai
category: ai
name: LocalAI
version: "v3.12.1"
version: "v4.0.0"
tagline: Drop-in OpenAI replacement
description: >-
LocalAI is the free, Open Source OpenAI alternative. LocalAI act as a drop-in replacement REST API that's compatible with OpenAI API specifications for local inferencing.
@ -15,17 +15,21 @@ description: >-
Before running a model, make sure your device has enough free RAM to support it. Attempting to run a model that exceeds your available memory could cause your device to crash or become unresponsive. Always check the model requirements before downloading or starting it.
releaseNotes: >-
This is a patch release fixing incompatibilities with Qwen 3 Coder, along with stability improvements and minor enhancements from v3.12.0.
This is a major release that transforms LocalAI into a complete AI orchestration platform with a fully rewritten React UI, native agentic capabilities, and many new backends.
Key highlights include:
- Fixed Qwen 3 Coder incompatibilities via updated llama.cpp
- Multi-modal Realtime - Send text, images, and audio in real-time conversations
- Voxtral Backend - New high-quality text-to-speech backend
- Multi-GPU Support - Improved Diffusers performance with multiple GPUs
- Improved UI with dark/light theme variants and better navigation
- Security fix for SSRF vulnerability in content fetching endpoints
- Various realtime stability fixes for audio, image, and model handling
- Native Agentic Orchestration - Manage, create, and run AI agents with memory and skills via the new UI, plus the new Agenthub community hub for sharing agents
- Revamped React UI - Complete frontend rewrite for a faster, modern experience with Canvas mode for previewing code artifacts
- MCP Client-Side Support - Full Model Context Protocol support with MCP Apps, tool streaming, and a new LOCALAI_DISABLE_MCP option to disable it
- WebRTC Realtime - WebRTC support for low-latency realtime audio conversations
- New Backends - Added experimental MLX Distributed, fish-speech, ace-step.cpp, and faster-qwen3-tts
- Data Separation - New LOCALAI_DATA_PATH env var to separate persistent data (agents, skills) from configuration
- Model RAM Warnings - Visual warnings when a model's storage size exceeds available system RAM
⚠️ Breaking Changes
- HuggingFace backend support has been removed
Full release notes can be found at https://github.com/mudler/LocalAI/releases