To enable the AI Assistance feature on the self-hosted Trendz application, or to not be related to the provided by cloud token limits, it is possible to configure your own large language model.
To enable/configure your own model, it’s necessary to go: Setting -> General -> AI Assistant
To use the own model it’s necessary to enable Use own model. After enabling the own model it’s possible to choose an AI provider and model. Additionally, it’s necessary to enter API key or Access Key and Secret Key (related to the chosen provider).
⚠️ Note: each AI Assistance request uses around 10000 input tokens and 1000 output tokes. Token usage is related to the difficulty of request and provided topology size.
Currently, Trendz could be integrated with the following models:
- OPEN_AI
- AMAZON_BEDROCK
Google (Recommended):
Please refer to the Google API key documentation to find out how to create an API key for Google models.
Supported models:
- Gemini-1.5-pro
- Input (1m) token price: 1.25$
- Output (1m) token price: 5$
- Gemini-1.5-flash
- Input (1m) token price: 0.075$
- Output (1m) token price: 0.3$
- Gemini-2.0-flash (Recommended - cost efficiency, average performance)
- Input (1m) token price: 0.1$
- Output (1m) token price: 0.4$
- Gemini-2.0-flash-lite
- Input (1m) token price: 0.075$
- Output (1m) token price: 0.3$
Open AI
Please refer to the Open AI API key documentation to find out how to create API keys for Open AI models.
Supported models:
- Gpt-4o (Recommended - high performance)
- Input (1m) token price: 2.5$
- Output (1m) token price: 10$
- Gpt-4o-mini (Recommended - cost efficiency)
- Input (1m) token price: 0.15$
- Output (1m) token price: 0.6$
- o3-mini (Recommended - cost efficiency, high performance but slow speed)
- Input (1m) token price: 1.1$
- Output (1m) token price: 4.4$
Amazon Bedrock
Please refer to AWS API key documentation to find out how to create Access key IDs and Secret access keys.
Supported models:
- Amazon (Recommended)
- Amazon Nova Lite (Recommended - cost efficiency)
- Input (1m) token price: 0.06$
- Output (1m) token price: 0.24$
- Amazon Nova Pro
- Input (1m) token price: 0.8$
- Output (1m) token price: 3.2$
- Amazon Nova Lite (Recommended - cost efficiency)
- Anthropic
- Claude 3.5 haiku
- Input (1m) token price: 0.8$
- Output (1m) token price: 4$
- Claude 3.5 sonnet (Recommended)
- Input (1m) token price: 3$
- Output (1m) token price: 15$
- Claude 3.7 sonnet
- Input (1m) token price: 3$
- Output (1m) token price: 15$
- Claude 3.5 haiku
- Meta Llama
- Llama 3.1 Instruct (70b)
- Input (1m) token price: 0.72$
- Output (1m) token price: 0.72$
- Llama 3.2 Instruct (90b)
- Input (1m) token price: 0.72$
- Output (1m) token price: 0.72$
- Llama 3.3 Instruct (70b) (Recommended)
- Input (1m) token price: 0.72$
- Output (1m) token price: 0.72$
- Llama 3.1 Instruct (70b)
- Mistral AI
- Mistral Large (24.02)
- Input (1m) token price: 3$
- Output (1m) token price: 12$
- Mistral Large (24.02)
⚠️ Note: to save any changes in the configuration it’s necessary to press the “Save” button.
Next Steps
-
Getting started guide - These guide provide quick overview of main Trendz features.
-
Installation guides - Learn how to setup ThingsBoard on various available operating systems.
-
Calculated Fields - Learn about Calculated fields and how to use them.
-
States - Learn how to define and analyse states for assets based on raw telemetry.
-
Prediction - Learn how to make forecasts and predict telemetry behavior.
-
Filters - Learn how filter dataset during analysis.
-
Available Visualizations - Learn about visualization widgets available in Trendz and how to configure them.