Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

jev(): a PostgreSQL extension for natural-language queries

X post showing a jev() PostgreSQL function that filters rows by natural-language conditions with no index or embeddings.

Discussiondiscussion
Open source ↗Live demo ↗

This summary is based on how curated lists describe the project; its README was not available to review.

Overview

This X post shows jev(), a PostgreSQL extension that adds a single SQL function for natural-language filtering. A query such as WHERE jev(people, 'could work from home') asks Jev to judge each row against the condition, so a whole table can be searched by meaning without building an index or embeddings.

It illustrates a pattern of putting a yes/no decision model directly in the database query path. Per-row model calls imply cost and latency that scale with the number of rows scanned, so it suits filtered or modest tables. Only the list summary was available; no repository, installation details or performance figures were found.

What the decision model decides

TypeQuestionAnswers
Yes / noDoes this row satisfy the natural-language condition (e.g. could work from home)?(inferred)yes · no

How it works

  1. Write SQL with jev(row, 'condition') in a WHERE clause
  2. Extension sends each row and the condition to Jev
  3. Rows above a threshold are returned

Strengths

  • Natural-language filters in plain SQL
  • No index or embeddings to maintain

Caveats

  • Cost and latency grow with rows scanned
  • Social post only; code and details not available

At a glance

Type
Discussion
Status
discussion
Runtime
PostgreSQL extension

Use cases and integrations

How curated lists describe it

  • X: a single SQL function that searches a whole database in natural language with no index and no embeddings, e.g. `WHERE jev(people, 'could work from home')`.

    yibie/awesome-jev

Frequently asked questions

What does jev() do?

It lets a SQL WHERE clause filter rows by a natural-language condition judged by Jev.

Does it need embeddings?

No, it judges rows directly with no index or embeddings.

Is it available to install?

The source is an X post; no repository was linked in the list.

Related projects

  • graphs-and-jev

    Accenture thought piece and site on stack migration with Google Cloud, placing Jev as the System 1 layer that routes, classifies and gates.

    MCP serverdiscussionRust
  • Databricks AI Decide

    Databricks blog post introducing ai_decide, an AI Function that classifies, scores and chooses over governed data.

    ArticleproductionSQL
  • A DuckDB extension for row classification

    DuckDB extension that labels rows in CSV, Parquet or DuckDB tables with Jev, reported at about ten seconds per thousand rows.

    Discussionprototype
  • anofox-decide

    DuckDB extension adding decide_probability, decide_choice and decide_score SQL functions backed by Jev or local open decision models.

    PluginbetaC++
  • dbt\_jev

    dbt package whose macros run Jev Choice, Noul and Score over SQL values in DuckDB and ClickHouse models.

    ProjectbetaSQL (Jinja)
  • duckdb-jev

    DuckDB extension that asks Jev about every row in SQL and returns answers as ENUM, DOUBLE or STRUCT columns.

    PluginbetaC++