LibreLM Playground

TRAINED ON A LAPTOP GPU

Intelligent Dynamics
Libre Language Model.

Try one of our lightweight language models in your browser.

Jump to the playground
Not loaded

This model runs on your device.
The model files download to this device. Your passages are not sent to an AI server.

LOCAL INFERENCE

Text completion

PRETRAINED · NOT A CHATBOT

Pick a starting point or write a few sentences, then press Generate. The model continues your passage. It has not been tuned for chat.

PICK A STARTING POINT

Choose an example to fill in the passage.

Prewritten openings. Your settings stay unchanged. The model generates only the continuation.

Load the model to count tokens

Continuation

The continuation will appear here.
Choose a starting point while the model prepares.

FIRST TOKEN

-After the model is loaded

DECODE SPEED

-On this device · excludes first token

GENERATED

-Tokens in the selected continuation

About this model

Model size and browser precision are read from the loaded export. A lower-precision version may lose quality, especially without further tuning.

Speed depends on your browser, available memory, and GPU support. Some phones can run this model, but we have not tested every device. It can produce errors, repetition, or unrelated text. Search compares continuations without checking their facts.

Local inference. The app processes passages and notes on your device. It has no remote inference endpoint or prompt logging. The website host may log file downloads, but those requests do not contain your text.