Skip to content
Papeliz
Use App

Bring your own AI provider

Papeliz’s assistant speaks the OpenAI-compatible format. Pick a preset, paste your API key, and the assistant talks straight from your device to that provider. A preset only fills in the base URL; you can change it, and the model, at any time.

Recommended

Muse Spark · Meta Model API

Papeliz’s assistant is built first around Meta’s Muse Spark models, and the Meta Model API is the preset selected out of the box. Long context lets the assistant work across many notes at once, and the API accepts calls straight from the app.

Base URL
https://api.meta.ai/v1
Models
muse-spark-1.3, muse-spark-1.1
API key
From the Model API dashboard

Other hosted providers

  • OpenAI

    OpenAI’s hosted models.

    https://api.openai.com/v1

  • OpenRouter

    Many hosted models behind one key.

    https://openrouter.ai/api/v1

  • Groq

    Fast hosted open-weights models.

    https://api.groq.com/openai/v1

  • Together

    Hosted open-weights models.

    https://api.together.xyz/v1

  • Fireworks

    Hosted open-weights models.

    https://api.fireworks.ai/inference/v1

  • Google Gemini

    Gemini models through Google’s OpenAI-compatible endpoint.

    https://generativelanguage.googleapis.com/v1beta/openai

On your own computer

Run models locally and nothing leaves your machine. Papeliz connects to the local server from your device, with no API key needed.

  • Ollama

    Models running on your computer. Allow Papeliz’s address in OLLAMA_ORIGINS first.

  • LM Studio

    Models from LM Studio’s server on your computer.

Any other provider

Choose Custom and type the base URL of any OpenAI-compatible API, hosted or running on your own machine.

Where your key goes

Your key is stored only on your device and sent only to the provider you pick. There is no Papeliz server in between. Read the privacy details or the guide to choosing a provider.