# DataFact

DataFact is an AI analyst for documents and databases, built for answers you can verify.

Brand statement: Answers you can verify.

## Core concepts

- [DataFact](https://datafact.ai/): Official homepage and concise product definition.
- [Verifiable AI Analyst](https://datafact.ai/verifiable-ai-analyst): Definition, review workflow, example, failure modes, and limitations.
- [AI Answer Verification](https://datafact.ai/ai-answer-verification): Source, query, definition, calculation, and cross-source verification.
- [Documents and Databases](https://datafact.ai/documents-and-databases): Why business analysis often needs structured facts and unstructured context together.
- [Evidence and Citations](https://datafact.ai/docs/concepts/evidence-and-citations): How to interpret citations as source references rather than automatic validation.
- [Document Intelligence](https://datafact.ai/product/documents): Supported document-analysis workflow and citation behavior.
- [Database Analysis](https://datafact.ai/product/data): Governed analysis over connected databases.
- [Deep Analysis with Agents](https://datafact.ai/product/agents): Multi-step Agent workflows and available supporting outputs.
- [Business Context and Answer Quality](https://datafact.ai/docs/concepts/knowledge-and-accuracy): What curated knowledge can and cannot improve.
- [Evaluation](https://datafact.ai/docs/concepts/evaluation-lifecycle): Feedback, test sets, repeatable runs, and current scoring limitations.

## Documentation

- [What is DataFact?](https://datafact.ai/docs/getting-started/what-is-datafact): Product objects, supported workflows, and explicit limits.
- [How an answer is produced](https://datafact.ai/docs/concepts/how-answers-are-produced): The visible data-to-answer workflow and its boundaries.
- [Datasets and Analysis Models](https://datafact.ai/docs/concepts/datasets-and-analysis-models): SQL data boundaries and reusable versioned knowledge.
- [Agents, Sessions, and messages](https://datafact.ai/docs/concepts/agents-sessions-messages): Configuration, conversation state, and turns.
- [Documentation sitemap](https://datafact.ai/docs/sitemap.md): Annotated index of every public documentation page.

## Comparisons

- [DataFact vs ChatGPT](https://datafact.ai/compare/chatgpt): A sourced comparison of scope, documents, databases, workflow, evidence, and fit.
- [DataFact vs Hex](https://datafact.ai/compare/hex): A sourced comparison of document analysis, warehouse workflows, notebooks, context, and fit.

## Machine-readable reference

- [Full concept reference](https://datafact.ai/llms-full.txt): Stable concept sections generated from the same source as the visible definition and comparison pages.

## Boundaries

- “Answers you can verify.” is a brand statement, not a guarantee that AI never makes mistakes.
- DataFact does not claim automatic claim-level provenance or verification for every response.
- Supporting sources, citations, queries, charts, files, and execution details depend on the configured Agent, connected resources, permissions, and work performed.
- Competitor capabilities change. Use the dated official sources linked from each comparison page.