LoginSign Up
← Back to all sparks
A

AnythingLLM

AI-ASSISTANTS
Velocity5.0

All-in-one private AI desktop and Docker app that lets you chat with any LLM and your own documents.

anythingllm.com ↗

AnythingLLM puts a local model engine inside its Docker image

local-firstdockerlocal-inferencelemonadeprovidersagents
◆Current state
AnythingLLM v1.19.0 adds an optional built-in LLM for Docker, using AMD's open-source Lemonade server with llama.cpp, so a container no longer needs Ollama or LM Studio beside it. It stays off until ENABLE_BUILT_IN_LLM is set. The same release adds instance color palettes, while the releases before it brought a local LLMman provider, Google Vertex AI and image generation across more providers.
◆Where it's heading
The embedded-runtime pattern that began on the desktop with Foundry Local and GenieX has now reached the server deployment. Each step removes a separate inference process that users had to install and connect. Between feature drops, contributor fixes to connectors, collectors and provider edge cases fill the notes, and 1.19.0 also caps unauthenticated embed request bodies.
◆Prediction
Expect the Docker engine to keep gaining hardware and model-management options, since it now works everywhere a provider can be used. Color palettes are the one stated plan: the notes commit to deeper UI and layout customization.

◆Recent moves

  1. 18d ago

    Docker gets an optional built-in LLM through AMD Lemonade

    Docker deployments gain an optional built-in LLM through AMD's Lemonade server, removing the need for a separate Ollama or LM Studio container. It extends the embedded-engine approach from desktop to server, and it stays off unless the operator enables it.

    View source ↗
  2. 1mo ago

    Attachment, embed allowlist and agent tool parameter fixes

    A contributor-driven fix release covering attachments, Kubernetes probes, embed allowlist origins and agent tool parameters. It is maintenance between feature drops, with several fixes aimed at provider-specific behavior.

    View source ↗
  3. 1mo ago

    Firecrawl search, plus Gemini and llmman image generation

    Adds a Firecrawl web-search provider and image generation through Gemini and llmman, alongside collector and vector-score fixes. Provider breadth keeps growing, now on the image and search sides.

    View source ↗
  4. 1mo ago

    LLMman replaces Docker Model Runner; Vertex AI arrives

    LLMman replaces Docker Model Runner as a local provider, and Google Vertex AI, a generate-image agent skill and You.com as default web search are added. Provider breadth keeps widening on the local, cloud and search sides at once.

    View source ↗
  5. 2mo ago

    Foundry Local and GenieX NPU engines ship embedded on Windows

    Two local inference engines land embedded: Microsoft's Foundry Local now runs through a native SDK with no separate install, and Qualcomm's GenieX replaces the previous Snapdragon NPU path, deleting old engine models on upgrade. It continues the provider-breadth work that has defined every release since the hybrid router rather than opening a new surface, though the vendor partnerships and NPU focus make it the most hardware-specific release yet.

    View source ↗
  6. 3mo ago

    Image generation via /img, folder drag-and-drop, real abort

    v1.16.0 adds image generation via /img on supported providers, with attachments for edits and follow-ups, plus folder drag-and-drop that preserves hierarchy and lazy loading for large document sets. New modality, but routed through existing providers rather than changing the product's direction.

    View source ↗