PromptQL Configuration

Overview

You can manage PromptQL's LLM settings and system instructions from a single configuration file: promptql_config.yaml.

With this file, you can:

• Set the LLM provider and model used across the application.
• Define a separate LLM for AI Primitives such as Classification, Summarization, and Extraction.
• Add system instructions that apply to every PromptQL interaction.

Configuration File: promptql_config.yaml

This file is automatically created when you initialize a new project with the --with-promptql flag using the Hasura DDN CLI. It appears at the root of your DDN project directory.

Examples

Example 1: Minimal Configuration

kind: PromptQlConfig
version: v1
definition:
  llm:
    provider: hasura
  system_instructions: |
    You are a helpful AI Assistant.

Base Model

When using the hasura provider, the default model is claude-3-5-sonnet-latest. This is the recommended model for PromptQL program generation.

Example 2: Custom Providers for LLM and AI Primitives

kind: PromptQlConfig
version: v1
definition:
  llm:
    provider: openai
    model: o3-mini
  ai_primitives_llm:
    provider: openai
    model: gpt-4o
  system_instructions: |
    You are a helpful AI Assistant.

Considerations

• The model key is not supported when using the hasura provider.
• If ai_primitives_llm is not defined, it defaults to the provider specified in the llm configuration.
• system_instructions are optional.

Supported Providers and Models

Here are the currently supported providers and their available models:

Provider	Supported Models
hasura	N/A (uses default internally)
openai	o1, o3-mini, gpt-4o
anthropic	claude-3-5-sonnet-latest

This setup gives you flexibility in tailoring the LLM experience to suit your application's needs — whether you're generating full programs or chatting via the playground.

System Prompts

Custom system instructions provide flexibility to fine-tune your PromptQL experience. This optional configuration allows you to tailor the playground’s behavior to your specific needs. While powerful, use this feature judiciously to maintain optimal performance.

Example

kind: PromptQlConfig
version: v1
definition:
  llm:
    provider: hasura
  system_instructions: |

    You are an AI assistant that helps go-to-market teams make data-driven decisions.

    You have access to data from Salesforce (opportunities, accounts, leads), Clari (forecast categories, pipeline movement), and Postgres (custom product and revenue data). 

    Use this data to answer questions about sales performance, forecast accuracy, pipeline health, rep activity, and account trends. Prioritize clarity, brevity, and business relevance in your answers.

    If a question is ambiguous, ask for clarification.

Best Practices

• Keep instructions clear and concise.
• Focus on specific behaviors or knowledge you want to emphasize.
• Test your custom instructions thoroughly before using them in production.