What did Hlido score Zaturn?

Zaturn scored 58/100 (FADING) in Hlido's independent, hands-on review.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

Zaturn

Name: Zaturn review
Item: Zaturn
Rating: 58
Author: Hlido Editor

Research · tested 2026-09-05 · re-test due 2026-12-04 · by the Hlido desk, not the vendor

In short: Open-source, local-first 'chat with your data' co-pilot that gives AI models SQL tools across many data sources, usable as an MCP server or a Jupyter-like studio — a promising, honest single-maintainer project that's early and asks you to bring your own LLM.

5 PASS · 0 FAIL of 5 public-surface claims

Visit Zaturn → scorecard.json

Quick answer

Zaturn scores 58/100 (FADING) on Hlido’s independent, hands-on test (reviewed 2026-09-05). FADING (58): a clear, genuinely useful concept with a real MCP path, broad data-source support and a privacy-friendly local-first design — well clear of FLATLINE.

Hlido Editor · 2026-09-05

Zaturn's pitch is refreshingly concrete: instead of building dashboards, you point an AI at your data and ask questions, and Zaturn provides the tools that let the model run SQL so you don't have to. It runs locally, works either as an MCP server for any protocol-capable AI client or through a bundled studio (web) interface reminiscent of a notebook, and it connects to a solid spread of sources — PostgreSQL, MySQL, SQLite, DuckDB, ClickHouse, SQL Server, BigQuery, plus CSV and Parquet files. The local-first design is a real privacy advantage for people who don't want their data flowing to a hosted BI vendor, and the 'bring your own LLM API key' model is honest about where the intelligence comes from. The tempering facts are all about maturity: this is an early, single-maintainer project (~75 stars), the 'with vibes' framing signals a young tool rather than a hardened one, quality of the generated SQL and analysis is untested at Tier-1, and there is no team/governance story on the surface. For an individual analyst or a curious builder it's an appealing, low-lock-in way to try agentic data analysis; for anything mission-critical it needs more track record than the surface shows.

Why FADING

FADING (58): a clear, genuinely useful concept with a real MCP path, broad data-source support and a privacy-friendly local-first design — well clear of FLATLINE. Held here because it is early and single-maintainer (~75 stars), the generated-SQL/analysis quality is untested, it requires the user's own LLM key, and there's no team/governance or reliability evidence on the public surface.

Public-surface checklist

PASS Homepage loads (required)
PASS Primary value prop (required) — 'Your open source, AI-powered data analysis co-pilot — Just chat with your data'
PASS Cta present (required) — 'Install'
PASS Pricing or access — Open-source, free to run locally; user supplies own LLM key
PASS Evidence or demo — Supported data sources and usage examples shown; demo video linked in README

What it does well

Concrete value: give an AI SQL tools so you query data by asking, not by building dashboards
Works as an MCP server for any protocol-capable client, or via a bundled studio UI
Broad data-source support (Postgres, MySQL, SQLite, DuckDB, ClickHouse, SQL Server, BigQuery, CSV, Parquet)
Local-first / runs on your PC — a real privacy advantage over hosted BI
Honest 'bring your own LLM key' model

What it fails at

Early, single-maintainer project (~75 stars) — limited track record
Quality of generated SQL and analysis is untested at Tier-1
Requires the user's own LLM API key with the studio interface
No team, governance or reliability story on the public surface

Best for

Individual analysts and builders wanting local-first, low-lock-in data Q&A
MCP users who want to give a Claude/other client SQL access to their databases
Privacy-conscious users who want to keep data off hosted BI tools
People comfortable supplying their own LLM key

Not recommended for

Teams needing a hardened, supported, multi-user BI platform
Mission-critical analytics that require vendor SLAs and a track record
Non-technical users who can't run a local package or manage an LLM key

Pricing & access

Pricing findable on the public surfacePASS Open-source, free to run locally; user supplies own LLM key (tested 2026-09-05)

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-09-05.

Compared to

Julius local-first-open-source
Julius is a polished hosted data-analysis product; Zaturn is a local-first, open-source, bring-your-own-LLM tool with an MCP path. Choose Julius for managed convenience, Zaturn for locality, low lock-in and agent integration.

Agent relevance

API CLI MCP Behavioral-testable

Runs as an MCP server exposing SQL tools to an AI client, or standalone via a studio UI. Directly agent-drivable and behaviourally testable; analysis quality not exercised at Tier-1.

Agent-friendly score: 7/10

Evidence

Gives AI models tools to run SQL so the user doesn't have to — source (2026-09-05) verified
Usable as an MCP server or a Jupyter-like studio web interface — source (2026-09-05) verified
Connects to BigQuery, ClickHouse, DuckDB, MySQL, Postgres, SQLite, SQL Server, CSV, Parquet — source (2026-09-05) verified
Runs locally; requires the user's own LLM API key (studio) — source (2026-09-05) verified

scorecard.json · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-12-04

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/kdqed-zaturn.svg)](https://hlido.eu/check/?agent=kdqed-zaturn)

HTML

<a href="https://hlido.eu/check/?agent=kdqed-zaturn"><img src="https://hlido.eu/badge/kdqed-zaturn.svg" alt="Hlido trust score"></a>