Plugin Marketplace
Sign in

PrintAIgotchi

TrustedLukáš BrzákNo votesgithub.com/lukas-brzak-prusa/print-ai-gotchi
  • Three characters to pick from: Nozzle the grumpy workshop cat, Bolt the sports-commentator robot and Blobby the zen filament blob.
    Three characters to pick from: Nozzle the grumpy workshop cat, Bolt the sports-commentator robot and Blobby the zen filament blob.
  • Ask the pet about your printers. It answers from what it sees, using a small model on your own server or a provider you choose.
    Ask the pet about your printers. It answers from what it sees, using a small model on your own server or a provider you choose.

A pet that watches your printers, tells you what needs you, and chats with a small AI model running on your own server.

This extension is native code that runs on your Connect Local server, with the permissions you grant it at install. It is not sandboxed. Prusa reviewed this release before publishing it; install extensions only from authors you trust.

Connects to
  • github.comfrom the server

    Downloads the pinned llama.cpp build when a person installs a local model.

    Sends: Only the request for the file.

  • release-assets.githubusercontent.comfrom the server

    GitHub serves the llama.cpp build from there after a redirect.

    Sends: Only the request for the file.

  • huggingface.cofrom the server

    Downloads the pinned model a person installs on the pet's page.

    Sends: Only the request for the file.

  • *.aws.cdn.hf.cofrom the server

    Hugging Face serves the model from there after a redirect.

    Sends: Only the request for the file.

  • api.anthropic.comfrom the server

    Asks Claude, when the brain is Anthropic Claude.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • api.openai.comfrom the server

    Asks OpenAI, when the brain is OpenAI.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • generativelanguage.googleapis.comfrom the server

    Asks Gemini, when the brain is Google Gemini.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • openrouter.aifrom the server

    Asks OpenRouter, when the brain is OpenRouter.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • api.groq.comfrom the server

    Asks Groq, when the brain is Groq.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • api.mistral.aifrom the server

    Asks Mistral, when the brain is Mistral.

    Sends: The API key, the pet's character, the state of the printers the person may see, and the conversation.

  • the server you set in base_urlfrom the server

    Asks the server at this address, or at the one a person enters on the pet's page for Ollama, LM Studio or another server, when page_setup is on.

    Sends: The API key if one is set, the pet's character, the state of the printers the person may see, and the conversation.

What's inside 0.2.0
  • module.yamlThe manifest: what the extension is and what it asks for
  • print-ai-gotchi_linux_amd64The program for linux amd64
  • print-ai-gotchi_linux_arm64The program for linux arm64
  • print-ai-gotchi_darwin_arm64The program for darwin arm64
  • print-ai-gotchi_darwin_amd64The program for darwin amd64
  • ui/The browser bundle the Connect web app loads
  • checksums.txtThe SHA-256 of every file above
  • checksums.txt.sigThe author's signature over checksums.txt

A release may also carry LICENSE, README.md and CHANGELOG.md. Nothing else is accepted.

Versions
0.2.0latest .tar.gz

MIT · core >=0.1.0-beta.1 · sink · linux_amd64, linux_arm64, darwin_arm64, darwin_amd64 · 16599.0 KB · Nov 17, 2026

Added

  • The pet gets hungry when nothing prints: peckish after 12 hours, hungry after 24, starving after

    1. It says so, shows a hungry face, and thanks whoever feeds it with a print. GET /pet carries vitals with its hunger and how full it is.

  • The pet grows with the printing hours of the whole fleet: hatchling, youngster, grown-up, legend. It is drawn smaller in its shell at first and wears a crown as a legend.

  • A morning roast: every morning from roast_hour (8 by default, -1 turns it off) the pet sums up the day before in its own character. It counts only, and names no printer.

  • The stage shows a hunger meter and a growth meter.

  • A model download can be cancelled on the Brain tab, or with DELETE /brain/download. Nothing of it stays on disk.

Changed

  • Downloading a model no longer switches the pet to it at once. The pet keeps its brain while the model downloads and moves to it once it is on disk, unless somebody chose another brain meanwhile. A cancelled or failed download leaves the pet able to talk.

  • Installing, removing and cancelling a model are refused when page_setup is off, like choosing a brain already was.

MIT · core >=0.1.0-beta.1 · sink · linux_amd64, linux_arm64, darwin_arm64, darwin_amd64 · 16515.3 KB · Nov 17, 2026

Fixed

  • Installing a model on the Brain tab works again. In 0.1.1 the first download stopped the module, and a download while the local brain ran failed.

Changed

  • llama-server gets only the environment it needs: its library path, PATH, HOME and TMPDIR.

Comments (0)

Sign in to comment.

No comments yet.