Dataherald/dataherald

Health v2 · maintenance only

Adoption not part of the score

3646 stars · 261 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Dataherald is an open-source natural language-to-SQL engine for enterprise question answering over relational databases, exposing an API that answers plain-English questions with generated SQL. It ships as a monorepo with an engine, an enterprise API layer with auth, an admin console, and a Slackbot, deployable via Docker Compose.

Use cases

• let business users query a data warehouse in plain English without an analyst
• add natural language Q&A over a production SQL database to a SaaS app
• build a ChatGPT plugin on top of proprietary relational data
• convert natural language questions into SQL queries with an LLM
• set up a Slack bot that answers data questions from a channel
• fine-tune text-to-SQL generation for a specific database schema

When to choose

• you need a self-hosted, API-first NL-to-SQL engine over relational data
• you want enterprise features like organizations, users, and authentication around text-to-SQL
• you want a GUI admin console and Slack integration out of the box
• you need fine-tuning support to improve SQL generation accuracy on your schema

When to avoid

• you need a lightweight library to embed in an existing app rather than a multi-service deployment
• your data is not in a relational/SQL database
• you need a no-code BI dashboard rather than a question-answering API
• you require actively maintained software - the latest release dates to mid-2024

Facets

service · maturity maintenance

rag llm-inference api-framework database chatbot databases large-language-models data-science apis self-hosted python self-hosted cross-platform text-to-sql nl2sql text-to-query data-warehouse business-intelligence enterprise slackbot admin-console retrieval-augmented-generation docker web-server

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Dataherald/dataherald")