Dataherald vs Savvina AI
Dataherald
github.com
Savvina AI
savvina.aiSide by side
| Figure | Dataherald | Savvina AI |
|---|---|---|
| Price from | — | €39/mo |
| Free plan | — | Yes |
| Free trial | — | — |
| Open source | — | — |
| Web | Yes | Yes |
| API | Yes | Yes |
| Self-hosted | Yes | Yes |
| Plans published | 0 | 5 |
| Spec score · AI SQL Generators | 7.7 (#10) | 8.0 (#6) |
| Credentials | The quick start requires an OpenAI API key, an organization ID, and an encryption key before starting the engine. | Database passwords and API keys are Fernet-encrypted at rest, with plaintext decrypted into memory when a query is dispatched. |
| Deployment | The quick start describes self-hosting the engine with Docker, with Mongo running in a separate container for application data. | All plans run on the customer’s own infrastructure via Docker, and the pricing page says Savvina does not touch customer data. |
| Intended users | The documentation is intended for developers building natural-language interfaces from structured data in their own projects and contributors to the engine. | The maker describes use by operations, data, engineering, finance, healthcare, HR, and other teams that need analytics while keeping data in-house. |
| Security | The quick start says database connection data is encrypted before storage in MongoDB using an ENCRYPT_KEY and recommends changing the default MongoDB username and password. | Savvina says query results are never sent to external models and that the LLM receives schema information rather than query results. |
Plans
Savvina AI
- CommunityFree
- Starter€39 a month
- Team€249 a month
- Business€1,299 a month
- EnterprisePriced on request
In their own words
Dataherald
Dataherald is an open-source natural language-to-SQL engine that exposes an API for asking questions about relational data in plain English. Its stated uses include helping business users query warehouses, embedding data Q&A in SaaS products, and creating ChatGPT plug-ins for proprietary data. The project targets developers building natural-language interfaces or contributing to the engine. Self-hosting uses Docker, with MongoDB in a separate container for application data. Listed warehouse connections are PostgreSQL, BigQuery, Databricks, and Snowflake. Included SQL-generation approaches are LangChain SQL Agent, LangChain SQL Chain, LlamaIndex SQL Generator, and Dataherald SQL Agent. A context store can include business and database details, validated question-and-SQL examples, and table information; active learning from usage is also described. An evaluator assigns confidence scores to generated SQL. The modular architecture allows implementations to be changed through .env configuration. Startup requires an OpenAI API key, organization ID, and encryption key. Connection data is encrypted before Mongo storage. The repository uses Apache-2.0 licensing.
Savvina AI
Savvina AI turns plain-English questions into SQL and returns database results within the customer’s own environment. It supports follow-up queries, validates SQL as read-only, and adds LIMIT clauses automatically. Results can appear in eight chart types, from bar and line charts to gauges and numeric displays. The product lists 22 data sources, including PostgreSQL, BigQuery, Snowflake, MongoDB, Amazon S3, and Google Cloud Storage. Customers bring their own LLM provider; listed choices include Anthropic, OpenAI, Google, Ollama, OpenRouter, and vLLM. Plans run on customer infrastructure through Docker. Savvina says the LLM receives schema details rather than query results; users can exclude schemas, tables, or columns from prompts, and sensitive values are masked in results. Database passwords and API keys are encrypted at rest. The free Community plan is limited to one user and PostgreSQL and MySQL. Paid plans start at €39/mo, with no free trial. Enterprise pricing is custom and includes multiple deployments and air-gapped licenses.