What AI Capabilities Does Lemonade Support?

Updated

Lemonade covers six capability areas in one local runtime: chat, vision, image generation, speech, transcription, and embeddings. It is designed to run local-first on all major platforms, while also being able to route requests to servers and cloud APIs on demand. If your question is whether you can handle text, image, and voice needs without stitching together separate tools, the answer is yes — with the caveat that "on demand" routing means some capabilities may depend on what you connect beyond your own machine.

The six capability areas

Lemonade describes its runtime as "omni modal," meaning a single install gives your app access to all of the following:

Capability What it covers
Chat Conversational text generation for apps and agents
Vision Image understanding as an input modality
Image Image generation
Speech Spoken output
Transcription Converting speech to text
Embeddings Vector representations for search, retrieval, and similar tasks

The practical value of this list is that you don't need a separate stack per modality. One runtime exposes all six, so an app that chats, reads images, and transcribes audio can share the same local backend.

How you actually call these capabilities

Capabilities are reachable through three surfaces, all installed together:

  • GUI — a graphical interface for interacting with Lemonade directly.
  • CLI — command-line access for scripting and quick tests.
  • API endpoints — programmatic access, plus an Embeddable SDK for adding native support or an embedded copy of Lemonade inside your own app.

There is also an API Reference and source on GitHub. The website's Quick Start is the entry point for learning how to add native support or embed Lemonade.

Local-first, with optional routing

Two design points shape what you can expect:

  1. Local-first on all major platforms. The runtime is built to run on your own computer, and installation is offered for Windows, Docker, Snap, Ubuntu, Arch, Fedora, Debian, and macOS.
  2. Routes to servers and cloud APIs on demand. If a capability or model isn't available locally, Lemonade can route the request outward. This is a flexibility feature, not a guarantee that every capability runs fully offline in every configuration.

Privacy and openness, as stated

Lemonade describes itself as free to use, transparent, and zero telemetry, with data staying where you put it and no upsell. It also states it is built in the open with a community, steered by open working groups and built alongside open-source engines. These are the project's own claims; the source is on GitHub if you want to verify them.

App compatibility

Because Lemonade works with the OpenAI API standard, it is integrated in many apps and works out-of-box with hundreds more. If an app already speaks the OpenAI API, pointing it at Lemonade is the usual path to using these capabilities inside that app.

What to check before committing

  • Which capabilities you need locally vs. routed. Confirm your target modality runs on your platform without a cloud fallback if offline operation matters to you.
  • Your platform. Installation paths exist for Windows, Docker, Snap, Ubuntu, Arch, Fedora, Debian, and macOS — pick the matching one.
  • Your app's API expectations. If it uses the OpenAI API standard, integration is the intended path; otherwise, the Embeddable SDK is the alternative.

For the current release highlights, the project states it ships a new release every week, so capability details are worth re-checking against the latest release notes.