Quick Keypoints
- Semantic layer translates plain-text prompts into structured database queries.
- Multiplayer workspace records team chat context to build a shared AI skill wiki.
- Secure authorization gates model data queries to respect user permissions.
- Pay-as-you-go pricing structured around Operational Language Units (OLUs).
What is PromptQL?
PromptQL (promptql.io) is an agentic semantic layer developed by Hasura that acts as a collaborative AI coworker. It encodes custom business logic and database structures, allowing AI agents to query enterprise data securely, while capturing daily chat context to maintain a shared team knowledge base.
PromptQL is an enterprise-grade semantic middleware that connects AI models with corporate databases. By defining a dictionary of business logic, it ensures LLMs interpret queries (such as customer churn or recurring revenue) accurately. Teams interact in chat panels where the model updates documentation and runs calculations.
Who Needs PromptQL?
Data engineering teams, product managers, and business analysts needing secure, context-aware AI access to data.
Primary Use Cases
- Streamlining team workflows, task assignments, and calendar scheduling.
- Summarizing long documents, meetings, and research papers.
- Analyzing data and tracking project progress metrics.
Important Features
- Agentic Semantic Layer: Translate complex SQL schemas into natural language definitions.
- Shared Wiki Chat: AI coworker tracks conversation history to compile team guides.
- Data-Level Authorization: Built-in permission gates prevent unauthorized data access.
- Multi-Step Query Plans: Breaks complex analysis tasks into sequential logic steps.
Current Updates About PromptQL
- PromptQL integrated Hasura metadata engines and expanded Slack integrations.
- Current Version: PromptQL Core v2.0
Pricing Plans
| Plan | Price |
|---|---|
| Pay-as-you-goOperational Language Unit usage tracking, unified LLM keys, team dashboard | .20/OLU |