MLC Ollama Studio

If you need AI images for real work rather than experiments, you don't want to wire node graphs. MLC Ollama Studio is the desktop application for professionals who have to produce images in series, organise them and find them again: one batch recipe delivers a whole series with a single click, folders and tags keep hundreds of motifs searchable, and every image keeps its prompt for the next variation.

Made for: manuals and technical documentation · websites and product pages · advertising and social media · training and presentation material.

Setup: you need an image server (Mac: Z-Image, Windows: MLC SD); image analysis, style extraction, prompt improvement and translation are optional. Which models for what – see Setup & Models.

Cost-effective and private: MLC Ollama Studio works with your own local AI – Ollama or an OpenAI-compatible image server on your network. No subscription, no per-image fees, and your motifs never leave your network.

⚠️ Important notice about Ollama versions (as of September 2026) Starting with Ollama v0.32.6, Ollama has temporarily removed its experimental image generation. Newer versions reject image requests with the error image generation models are not currently supported (HTTP 400) – even though image models such as x/z-image-turbo still show up in the model list. For image generation via Ollama, please run Ollama v0.32.5 on your server and disable automatic updates. Text, vision and translation models keep working with current Ollama versions. From version 0.16.3 on, MLC Ollama Studio can generate images through an OpenAI-compatible image server instead (e.g. txt2zimage) – Settings → Server & Models → Image API.

Key Features

Model Variety & Control

Leverage any Ollama-compatible server to access a wide range of image generation and analysis models. Whether you are using Stable Diffusion, Llava, or specialized Prompt Improvement models, you have full control over the seeds, inference steps, and resolutions.

Batch Recipes & Showcase

A prompt with variables such as $emotion$ and a list of values turns into a whole series of images with one click. From version 0.16.3 on, MLC Ollama Studio keeps these batch recipes including their values and ships five ready-made showcase recipes – emotions, body language, art styles, photography and eras. They are exactly the prompts behind our Showcase: pick a recipe, press Generate, done.

Visual Intelligence

Unlock professional workflows with advanced visual analysis:

• Style Extraction: Analyze an existing image to extract its artistic essence and apply it to new prompts.
• Image-to-Prompt: Use multimodal models to automatically generate descriptive prompts from any image.
• Expert Translation: Utilize specialized translation models (like mlc-deenes-expert) to refine prompts across languages.

Global Accessibility & Native Localization

MLC Ollama Studio is built for a global community. We have moved to a fully externalized translation system (i18n) to ensure a seamless experience regardless of your primary language:

• Native Support: Full, high-quality translations for English, German, and Spanish are available out of the box.
• Future Ready: French support is currently in development and will be available in an upcoming release.
• Community Driven: Our localization system is designed for easy expansion, allowing for rapid integration of new languages.

Local Image Management

Keep your creations organized with the built-in Organizer. Group images into folders, tag them for easy retrieval, and manage your generation history with a clean, grid-based interface.

Performance & Security

By utilizing local Ollama servers, your data remains secure and your workflows remain private. The application is optimized for performance on macOS, Linux, and Windows, featuring a responsive Svelte-based frontend.