Dataherald
Open-source natural-language-to-SQL engine for enterprise Q&A.
About
Dataherald is an open-source (Apache-2.0) natural-language-to-SQL engine built for question answering over relational databases. You stand up an API in front of your warehouse or production DB so business users can ask questions in plain English; under the hood it does database scans to learn tables and columns, keeps golden question-to-SQL records for few-shot prompting, and applies database-level instructions to steer the LLM. The repo ships an engine, an enterprise API layer with auth, an admin console, and a Slack bot.
Its core modules are replaceable: the SQL generator, vector store for context, persistence DB, and an evaluator that scores generated SQL. The idea is solid and the architecture is well documented. The big caveat: the main repo's last commit was July 2024, over two years ago. Treat it as a proven reference design for NL-to-SQL rather than a maintained product, and check the fork network before deploying.
What it can do
- Natural-language-to-SQL REST API
- Automatic database schema scanning
- Golden records for few-shot prompting
- Database-level SQL instructions
- Replaceable SQL generator, vector store, and evaluator modules
- Admin console for configuration and observability
- Slack bot for chat-based querying
Setup
- Clone github.com/Dataherald/dataherald.
- Configure your database connection and LLM provider (e.g. OpenAI).
- Run the engine with Docker Compose.
- Run a database scan, add golden records and instructions, then query the REST API in plain English.
Prerequisites
Docker; a relational database; an LLM API key (OpenAI by default).
Use cases
Works with
Trust score — 6/10
Scored 6/10 (TIER metrics, metrics-v1). Evidence: 3,646 GitHub stars and 263 forks (live API check), Apache-2.0, well-documented modular architecture cited in agentic-analytics references. Strengths: strong community interest, clean design with replaceable modules, genuinely useful reference architecture. Docked for: (1) last commit 2024-07-24, over two years stale, so maintenance is effectively dead; (2) enterprise-production readiness claims predate the staleness and are untested. License: Apache-2.0.
★ 3,646 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 7.0/10
- Maintenance · recency, releases, issue load
- 3.0/10
- Openness · license + docs
- 8.0/10
- Manual review · sanity read + red-flag check
- 5.0/10
- · 3,646 GitHub stars, 263 forks (live API check)
- · Last commit 2024-07-24: over two years stale, maintenance effectively dead
- · Modular NL-to-SQL architecture well documented; cited in agentic-analytics references
- · Treat as reference design, not a maintained product
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.