FAIRplexica is an open-source AI assistant for research data management (RDM). It is based on Perplexica, uses the metasearch engine SearXNG to retrieve relevant RDM resources and local LLMs (via Ollama) to answer user questions. API providers like OpenAI, Groq, Anthropic, Google as well as custom API endpoints are available as well.

1. Ensure Docker is installed and running on your system.

2. Clone the FAIRplexica repository:

   git clone https://github.com/UB-Mannheim/FAIRplexica

3. Navigate to the project directory:

   cd FAIRplexica

4. Duplicate the sample.config.toml file and rename it to config.toml.

   • Fill in the empty fields depending on the LLM Providers you want to use; i.e. provide valid API keys for OpenAI, Anthropic etc.
   • Set your SearXNG URL
   • Set your Ollama URL if you have Ollama running and want to use it

     Note: You can change API keys after starting FAIRplexica inside the Settings dashboard.

   Example config.toml

   [GENERAL]
   ...
   GLOBAL_CONTEXT = "research data management"
   
   ...
   
   [ADMIN]
   # Provide a username for the admin account
   USERNAME = "admin"
   
   # Provide a secure password for the admin account 
   PASSWORD = "changeme"
   
   # Provide a base64 string with a length of at least 32 characters.
   # You can create your JWT_SECRET with a common shell command like: 
   # openssl rand -base64 32.
   JWT_SECRET = "replace-with-secure-secret" 
   
   ...
   
   [MODELS.OLLAMA]
   # If running Ollama models make sure the correct Ollama port is used
   # depending on your setup
   API_URL = "http://host.docker.internal:11434"
   API_KEY = ""
   
   ...
   
   [API_ENDPOINTS]
   # Make sure the correct SearXNG port is used depending on your setup
   SEARXNG = "http://host.docker.internal:4001/"

5. Change to the main project directory (containing the docker-compose.yaml) and execute:

   docker compose up -d

   Docker will pull the image, install everything and start the server. This may take a while.

6. Open http://localhost:3000/ in your browser, to check if everything works correctly.

7. Navigate to http://localhost:3000/admin and log into the Admin Settings with the USERNAME and PASSWORD you have set in section 4.

   • Inside the dashboard you can set your LLM and embedding models as well as API keys and your Ollama URL.