About

Project identity, constraints, and where to go next

GitHub Posit Connect Cloud Usage Notebooks

About

smol-doc-analyzer (package 1.0.0) explores whether a small, specialized, locally hosted pipeline can reach “good enough” accuracy for insurance document intake without sending claims documents to frontier APIs.

What you get

Dual inference paths: DICIE and the memo chain
Synthetic sample corpus + classical Random Forest baseline
Optional Discord front-end (Chloride)
This Quarto documentation portal (preview locally or publish to Posit Cloud)

Non-negotiables

No real insurer data

Training and demos use synthetic or public sources only. See Data Provenance.

Jump in

---
title: "About"
subtitle: "Project identity, constraints, and where to go next"
about:
  id: about-body
  template: trestles
  image: assets/logo.svg
  image-alt: "smol-doc-analyzer mark"
  links:
    - icon: github
      text: GitHub
      href: https://github.com/Exios66/smol-doc-analyzer
    - icon: cloud
      text: Posit Connect Cloud
      href: https://019f8d0a-2732-fecb-b056-1f69f4451c00.share.connect.posit.cloud
    - icon: book
      text: Usage
      href: usage.md
    - icon: journal-code
      text: Notebooks
      href: notebooks/index.qmd
---

::::: {#about-body}

::: {.hero-kicker}
About
:::

**smol-doc-analyzer** (package `1.0.0`) explores whether a small,
specialized, **locally hosted** pipeline can reach “good enough” accuracy for
insurance document intake without sending claims documents to frontier APIs.

### What you get

- Dual inference paths: [DICIE](docie_pipeline.qmd) and the [memo chain](architecture.qmd)
- Synthetic sample corpus + classical Random Forest baseline
- Optional Discord front-end (Chloride)
- This Quarto documentation portal (preview locally or publish to Posit Cloud)

### Non-negotiables

::: {.callout-important}
## No real insurer data

Training and demos use synthetic or public sources only.
See [Data Provenance](data_provenance.md).
:::

### Jump in

1. [Quick Start](quick-start.qmd)
2. [Pipeline hub](pipelines/index.qmd)
3. [Commands cheat sheet](reference/commands.qmd)

:::::