Evidence
AI NativeBusiness intelligence as code: build polished data products with SQL and markdown
MarketsData Visualization
Business intelligence as code: build polished data products with SQL and markdown
Published Oct 11, 2026
Sources:
Discussions referencing Evidence appeared in community forums influenced by Hacker News. Direct HN access was not consistently retrievable in this research session; however, community discourse indicates the product has been discussed in developer and build-in-public ecosystems.
Source:
Horizontal tool; likely adoption where data teams are software-oriented:
Get a comprehensive analysis of Evidence including market position, competitive landscape, adoption trends, and peer benchmarks.
Download full report →You require self-serve BI for non-technical business users If the objective is to enable non-technical analysts to iterate on dashboards without engineering involvement, tools such as Metabase, Tableau, or Power BI may be more appropriate.
You require extensively documented compliance artifacts upfront If procurement requires a public trust center, detailed compliance mappings, penetration test reports, or explicit data residency documentation, the vendor's public materials may be insufficient until additional documentation is provided.
You need extensive drag-and-drop modeling or a GUI semantic layer Evidence is SQL-first; organizations seeking a centralized semantic layer and GUI-based modeling may find it limiting.
Your team cannot absorb SDLC overhead for analytics If pull-request reviews, deployments, and code ownership are a poor cultural fit, a BI-as-code approach can slow operations.
Sources:
Sources:
Sources:
Source:
This limited third-party review volume is a constraint for buyers seeking extensive independent feedback.
Sources:
Vendor-published quotes on the Evidence site attribute positive statements to named roles at customer organizations. These vendor-selected quotes signal some customer satisfaction but are not independent evidence. Examples paraphrased from vendor material:
Interpretation: these vendor-provided quotes are useful as directional indicators of use cases and customer roles but should not substitute for independent reference checks.
Source:
Public community discussions show both interest and skepticism.
Examples (mixed sentiment):
Sources:
Vendor materials include customer names and quotes, and reference multiple organizations using the product. However, detailed, independently validated case studies with methodology and before/after metrics were not located in the accessible materials.
What is available (vendor-selected): named roles and organizations citing benefits.
What is missing for rigorous evaluation:
Sources:
Inferred fit based on product nature and community/OSS adoption patterns:
These are inferences; the vendor does not publish explicit ICP bands in accessible sources.
Evidence is an open-source business-intelligence-as-code framework for building and publishing data products (reports, narrative analytics, and embedded dashboards) primarily using SQL and Markdown. It generates a website or app from code rather than using a drag-and-drop BI interface. The product is intended for teams that prefer analytics artifacts to be managed like software: version controlled, reviewable, testable, and deployable.
Evidence’s core workflow is:
Sources:
Positioning: a code-first alternative to drag-and-drop BI tools. Evidence emphasizes narrative reporting, reproducibility, and integration with developer workflows (Git, VS Code, CI/CD) rather than ad-hoc exploration for non-technical business users.
Likely buyers/users:
Evidence's documentation states that users must know SQL; basic Markdown familiarity is also required. This is a functional constraint compared with traditional BI aimed at non-technical analysts.
Sources: