Run PapersGPT for Zotero in 1 Minute

Supercharge your research workflow with PapersGPT, the leading Zotero AI plugin.

Already have an API Key? If you possess a valid API key for OpenAI, Gemini, Claude, DeepSeek, OpenRouter, Qwen, Mistral, Kimi, GLM, or SiliconFlow, simply enter it directly into the PapersGPT model selection field. You can immediately start processing vast amounts of literature.

No API Key? No Problem. You can use PapersGPT for free instantly. Below are two methods to access powerful AI models without a paid subscription.

1. Free Cloud LLM API Services

Use these services to access hosted Large Language Models (LLMs) with free tiers.

OpenRouter

  • How to use: Register and create an API Key. Search for "Free" in the model box; any model tagged "Free" is cost-free.
  • Note: Availability is dynamic. Models like Grok 4.1 may be free during testing but require credits later.
  • Config: Input your OpenRouter API Key in PapersGPT as the picture below.
image

Groq

  • How to use: Register and create an API Key. Known for incredible inference speed.
  • Limits: Be aware of daily request limits. View Rate Limits.
  • Config: Select Customized in PapersGPT settings.
  • Config: API URL - https://api.groq.com/openai/v1/chat/completions in PapersGPT settings.
  • Config: Models must be the ones in Free Tier Limits.
imageimage

Mistral AI

  • How to use: Register and create a Mistral API Key at Mistral Official Website.
  • Benefit: Access to the free API tier (subject to rate limits).
  • Config: Input your Mistral API Key in PapersGPT settings.
image

Qwen

  • How to use: Register via the Qwen Landing Page.
  • Benefit: Receive a free usage quota for cloud-hosted Qwen models.
  • Config: set as Mistral AI.

GLM

  • How to use: Register at www.z.ai and generate an Key.
  • Benefit: Currently, the GLM-4.5-Flash model is free to use.
  • Config: set as Mistral AI.

SiliconFlow

  • How to use: Register at www.siliconflow.com.
  • Benefit: Receive a free quota upon registration.
  • Config: set as Mistral AI.

2. Local Free Open-Source Models

For maximum privacy and zero cost, deploy a permanently free Local LLM directly on your computer.

System Requirements

  • OS: Windows, macOS or Linux.
  • GPU: Graphics card required.
  • VRAM: At least 1GB free VRAM is required. If insufficient, there will be no local models for choosing.
  • Note: After selecting a local model, it will be automatically downloaded from Huggingface. Therefore, please maintain a stable network connection with Huggingface during the model download process.
  • Note: Once the model is downloaded, you can freely chat PDFs in Zotero, and you can even use it when there is no internet connection, such as on a plane or train.

Configure your local model as shown below:

image