OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Blendsql

Blendsql is an open-source Python query-language library that combines SQL with local language models to work across structured and unstructured data, with support for type constraints and constrained generation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Blendsql — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Blendsql when building a Python data workflow where SQL operations need to work alongside language-model reasoning over structured records or unstructured context.

Who it is for

  • Python developers building database-backed applications that need language-model processing alongside SQL.
  • Data and AI practitioners who want to query records with natural-language questions or map values to specified options.
  • Developers prototyping workflows that combine database tables with unstructured text context.

Documented uses

  • Combine SQL filtering, grouping or selection with language-model questions about records.
  • Map data to one of a defined set of options using an LLMMap query.
  • Provide unstructured text as context to an LLMQA query.
  • Constrain generated outputs by type or specified choices.
Before you choose

Requirements & limits

The repository is public and identifies the project as open source under the Apache-2.0 license. Its README lists Python 3.10–3.13 and provides an installation command using uv.

  • The documented interface is a Python library, so it suits programmatic workflows rather than a ready-made hosted app.
  • The README quickstart configures VLLM with a local base URL; the model service is part of that example’s setup.
  • The examples use pandas DataFrames as tables and demonstrate both question answering and option-constrained mapping.
  • The README includes a text-loading example that uses a URL as context.
  • The listed Python versions are 3.10, 3.11, 3.12 and 3.13.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Blendsql and prepare the documented Python setup

Install the library and configure the model-and-data objects shown in its quickstart.

Before you start

  • Python 3.10, 3.11, 3.12 or 3.13, as listed by the repository.
  • The README’s installation command uses uv.
  • The quickstart imports pandas and configures a VLLM model with a base URL at http://localhost:8000/v1/.
  • DataFrames and a model are passed to the BlendSQL connection.

Steps

  1. Install uv and Blendsql using the README command: pip install uv && uv pip install blendsql.

  2. Import pandas, BlendSQL and VLLM in Python.

  3. Configure the VLLM model with the model identifier and base URL shown in the quickstart.

  4. Create a BlendSQL connection with the required data tables and the model.

Expected outcome

A Python environment with Blendsql installed and a BlendSQL connection configured in the form shown by the quickstart.

Guide limitations

  • The README documents setup and configuration here; this guide does not verify that the model endpoint is running or execute a query.
  • The listed Python versions and the VLLM example do not establish compatibility with other versions or model backends.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
parkervg/blendsql · Official homepage title · retained in its source language
How we review sourcesSuggest a correction