Settings

Settings hold the configuration that controls how the Prometheux platform behaves for you: your authentication token, which LLM provider and model power AI features, embedding settings, and the limits that apply to your account. Configuration is user-scoped by default and is persisted on the server, so it follows you across sessions. The same values can be read and written through the Python SDK and the REST API — see the full Config reference for every function and endpoint.

Authentication token

The PMTX_TOKEN authenticates you against the platform. You can set it through an environment variable or through configuration:

import os

os.environ['PMTX_TOKEN'] = 'pmtx_token'

import prometheux_chain as px

px.config.set("PMTX_TOKEN", "pmtx_token")

LLM and model selection

Prometheux uses LLMs to generate human-readable explanations, translations of Vadalog rules, graph rag, and LLM output validation tasks. By default, no LLM is configured. To enable an LLM, set the provider and supply an API key:

import prometheux_chain as px

# Choose a provider
px.config.set("LLM_PROVIDER", "OpenAI")

# Provide the API key
px.config.set("LLM_API_KEY", "your_api_key")

You can also select the model version and tune generation parameters:

px.config.update_config({
    "LLM_PROVIDER": "OpenAI",
    "LLM_API_KEY": "your_openai_api_key",
    "LLM_VERSION": "gpt-4o",
    "LLM_TEMPERATURE": 0.7,
    "LLM_MAX_TOKENS": 4000
})

Default configuration values

The following defaults apply unless you override them:

Key	Default Value	Description
`PMTX_TOKEN`	None	Prometheux authentication token
`LLM_API_KEY`	None	LLM API key
`LLM_PROVIDER`	OpenAI	LLM provider name
`LLM_VERSION`	gpt-4o	LLM model version
`LLM_TEMPERATURE`	0.50	LLM temperature setting
`LLM_MAX_TOKENS`	2000	Maximum tokens for LLM responses
`EMBEDDING_MODEL_VERSION`	text-embedding-3-large	Embedding model version
`EMBEDDING_DIMENSIONS`	2048	Embedding dimensions

Saving and loading configuration

User configuration is persisted on the server and can be saved or loaded at any time.

import prometheux_chain as px

# Save user-scoped configuration
px.save_user_config(config_data={"theme": "dark"}, scope="user")

# Load it back
config = px.load_user_config(scope="user")

The equivalent REST endpoints are POST /api/v1/users/save-config and GET /api/v1/users/load-config. See the Config reference for request and response details.

Usage and limits

Your account has usage limits for LLM and embedding calls. You can check the current usage and remaining quota:

GET /api/v1/users/usage-status

A typical response:

{
  "data": {
    "llm_usage": {
      "current": 150,
      "limit": 1000,
      "remaining": 850
    },
    "embedding_usage": {
      "current": 50,
      "limit": 500,
      "remaining": 450
    }
  },
  "message": "Usage status retrieved successfully",
  "status": "success"
}

Your role on the platform can be retrieved with GET /api/v1/users/get-role.

Learn more

Config reference — complete list of configuration functions, endpoints, and parameters
Python SDK — using prometheux_chain to manage configuration programmatically
REST API — HTTP endpoints for configuration and user management

Introduction

Get Started

Platform

Integrations

Examples

Vadalog Reference

Resources

Authentication token

LLM and model selection

Default configuration values

Saving and loading configuration

Usage and limits

Learn more

​Authentication token

​LLM and model selection

​Default configuration values

​Saving and loading configuration

​Usage and limits

​Learn more

Authentication token

LLM and model selection

Default configuration values

Saving and loading configuration

Usage and limits

Learn more