agent · data analysis

Wren AI

Wren AI is a data-analysis agent that turns natural language into governed SQL across 20+ data sources including BigQuery, PostgreSQL, ClickHouse and Amazon Redshift, running tools and scripts inside an isolated environment and describing its design as sandboxed, multi-step reasoning rather than chat. Agents query, chart, extract from PDFs, build dashboards and save skills. Its governance layer is the strongest thing it publishes: row- and column-level scoping, with access enforced at query time across every caller, human or agent, and role-based access with full activity logs that the vendor says make it provable who saw what and which query produced which number — though no tamper-evidence mechanism and no retention period are described. It is deployable as multi-tenant cloud, private cloud or air-gapped on-premise, exposes REST endpoints for embedding, and supports MCP. The engine is described as open source with no licence named on the page checked. A customer testimonial claims up to an 80% query hit rate. Adoption figures on the page are not internally consistent: it claims 15,000+ data experts and analytics teams in one place and 50,000+ users in another, and the two are not reconciled. No certification, regulatory posture, AI liability cover, indemnity or outcome-linked payment term is published on the page checked.

www.getwren.aievidence retrieved Aug 9, 2026view as JSON all data analysis entries 16 alternatives

Reviewed12 or more fields evidenced, and scored by a human against the published rubric

16 fields evidenced · 43 with no public information. Every value below links to the document it came from and the date we checked it.

8 of those are quoted but not yet interpreted: we hold the source sentence and the date we read it, but nobody has written the answer to the question yet. Those rows show the quote and say so rather than repeating it back as an answer. Why these exist

9 of the evidenced values quote wording our own later capture of the same page no longer contains. Marked below, awaiting an editor.

Also appears in

AI agents that support the Model Context Protocol Open-source AI agents Self-hosted and on-premise AI agents AI agents and vendors that publish a self-reported performance figure

Trust gap

How this is scored

Runtime governance

2/3

Are scope, spend, and authority enforced while the agent runs?

Scope is enforced at execution time rather than advised: row- and column-level scoping is applied to every caller, and the page states access is enforced at query time across humans and agents alike. Short of 3 because no spend or authority ceiling is published and the enforcement is a vendor claim with no independent verification.

Counts Audit trail and Permission scopes, each cited below.

Audit trail

2/3

Is there a tamper-evident record of what it actually did?

Publishes a customer-reviewable record with attribution: "Role-based access with full activity logs. Provable who saw what, which query produced which number." Short of 3 because "provable" is asserted without a mechanism — no hash-chaining, no tamper-evidence claim, and no retention period.

Counts Audit trail, each cited below.

Compliance

0/3

Are compliance obligations attached per engagement?

No public evidence found. The page checked names no certification, attestation or regulatory posture. Its "Unified Data Policy" section describes access control (RLS and CLS), not a compliance position, and was rejected as compliance evidence in this audit.

How this gap gets closed →

Outcome settlement

0/3

Does payment depend on a verified result?

Payment is decoupled from result: a free open-source tier with a paid upgrade for governance, embedded API and enterprise support. No escrow, milestone release, clawback or outcome-linked term is published.

Counts Pricing model, each cited below.

How this gap gets closed →

Insurance & indemnity

0/3

Can the deployment be insured, and is the customer indemnified?

No public evidence found. The page checked publishes no AI liability cover for a customer deploying Wren AI, no named carrier, no policy limits, no customer indemnity and no liability cap.

How this gap gets closed →

Assurance

Insurance available: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Own liability cover: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Insurance carriers: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Coverage limits: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Indemnification: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Liability cap: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Audit trail: not yet interpreted
“Role-based access with full activity logs. Provable who saw what, which query produced which number.”
www.getwren.ai · checked Aug 9, 2026
Tamper-evident log: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Explainability: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Runtime governance: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Permission scopes: not yet interpreted
“Row & column-level scoping”
www.getwren.ai · checked Aug 9, 2026
Compliance certifications: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Regulatory alignment: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Outcome-based pricing: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Settlement mechanism: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Asset custody: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Dispute process: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
SLA terms: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Evaluation coverage: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Self-reported performance: A customer testimonial on the page claims "up to an 80% query hit rate"
“delivers up to an 80% query hit rate”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026

Agency

Autonomy level: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Human oversight: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Goal complexity: Multi-step reasoning in a sandbox, positioned against single-turn chat
“Agentic Design 01 Not just chat. Sandboxed, multi-step reasoning.”
www.getwren.ai · checked Aug 9, 2026
Action space: not yet interpreted
“Tools and scripts run inside an isolated environment. Agents query, chart, extract from PDFs, build dashboards, save skills.”
www.getwren.ai · checked Aug 9, 2026
Operating environment: not yet interpreted
“20+ data sources: BigQuery, PostgreSQL, ClickHouse, Amazon Redshift, and dialect-agnostic.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Initiative: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this

Safety

Safety evaluations: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Red teaming: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Safety policy: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Usage restrictions: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Model or system card: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Incident reporting: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Third-party evaluations: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Data handling: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this

Practicality

Pricing model: not yet interpreted
“Start free with open-source GenBI. Upgrade to full Agentic GenBI for governance, embedded API, and enterprise support.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Price point: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Availability: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Deployment options: not yet interpreted
“Multi-tenant cloud, private cloud, and air-gapped on-prem.”
www.getwren.ai · checked Aug 9, 2026
Integrations: 20+ data sources including BigQuery, PostgreSQL, ClickHouse and Amazon Redshift
“20+ data sources: BigQuery, PostgreSQL, ClickHouse, Amazon Redshift, and dialect-agnostic.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Supported regions: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Support model: "Enterprise support" named as a paid-tier feature; no response times or remedies published
“Upgrade to full Agentic GenBI for governance, embedded API, and enterprise support.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026

Foundation models

Base models: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Model provider: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Model swappable: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Open weights: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Fine-tuning: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Context window: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this

Ecosystem

Protocols supported: MCP; MDL, text-to-SQL and a CLI listed in the same capability row
“MDL · text-to-SQL · MCP · CLI · 20+ connectors”
www.getwren.ai · checked Aug 9, 2026
Tool use: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Multi-agent: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
API access: not yet interpreted
“Embedded AI API REST Build your own UX on the raw endpoints: natural language to governed SQL with execution, plus chart specs you render in your stack.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Open source: Engine self-described as open source; no licence named on the page checked
“The engine is open source. That's the difference.”
www.getwren.ai · checked Aug 9, 2026
Marketplace presence: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this

Impact

User base: Self-reported 15,000+ data experts and analytics teams. The same page separately claims 50,000+ users (see deployment_scale); the two figures are not reconciled on the page
“Trusted by 15,000+ data experts and analytics teams worldwide.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Deployment scale: not yet interpreted
“Trusted by 50,000+ users worldwide. Public companies across the US, EU, Middle East, and Asia.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
Target sectors: Manufacturing, healthcare, finance, retail and media
“Use cases in production across industries How companies across manufacturing, healthcare, finance, retail and media ship governed answers to humans and agents.”
Not in our latest capture. This quote was on the page when we filed it. We captured the same page again on Sep 17, 2026 and it was not there. The value above has not been re-checked by an editor yet. How we re-check citations
www.getwren.ai · checked Aug 9, 2026
High-risk domains: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Documented incidents: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this
Market recognition: no public information
www.getwren.ai · checked Aug 9, 2026 · source did not state this

Compared with

Side-by-side on the fields where Wren AI and the other entry are both on record and actually differ.