# achemnomaer/legal-draft-ai architecture

> Architecture diagram of the GitHub repository achemnomaer/legal-draft-ai, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-06). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/achemnomaer/legal-draft-ai
- Repository: https://github.com/achemnomaer/legal-draft-ai

## Overview

LegalDraft AI is a legal-document copilot: an operator uploads a document, reviews extracted fields, generates a cited draft, and can submit edits that inform later drafts. The backend extracts text and structured data, indexes chunks for retrieval, then drafts and optionally verifies grounded output; learned preferences feed subsequent drafting. The graph includes the documented frontend capabilities and provider options. Some subsystem links rely on README descriptions or SOURCE INDEX references because their implementation excerpts were not provided.

Main entry points: [`backend/generation/drafter.py`](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/drafter.py) (Grounded drafter), [`backend/ingestion/router.py`](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/router.py) (Ingest and documents), [`backend/main.py`](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/main.py) (FastAPI application), [`backend/llm/factory.py`](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/factory.py) (LLM provider factory)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_frontend["Review workspace"]
  node_upload_ui["Upload and review<br/>[page.tsx]"]
  node_document_view["Document view<br/>[page.tsx]"]
  node_draft_panel["Draft workspace<br/>[DraftPanel.tsx]"]
  node_draft_renderer["Citation renderer<br/>[DraftRenderer.tsx]"]
  node_diff_editor["Edit comparison<br/>[DiffEditor.tsx]"]
  node_preferences_ui["Preference viewer<br/>[PreferenceList.tsx]"]
  node_api_client["HTTP client<br/>[api.ts]"]
end

subgraph group_ingestion["Document ingestion"]
  node_ingest_route["Ingest and documents<br/>[router.py]"]
  node_text_extraction["PDF and OCR extraction<br/>[extractor.py]"]
  node_ocr["OCR processing<br/>[ocr.py]"]
  node_structured_extraction["Structured fields<br/>[structured.py]"]
end

subgraph group_retrieval["Search and grounding"]
  node_chunking["Document chunking<br/>[chunker.py]"]
  node_rag_store[("Vector index<br/>[store.py]")]
  node_hybrid_retrieval["Hybrid retrieval<br/>[hybrid.py]"]
  node_rag_route["Query endpoint<br/>[router.py]"]
end

subgraph group_drafting["Draft generation"]
  node_draft_route["Draft endpoint<br/>[router.py]"]
  node_drafter["Grounded drafter<br/>[drafter.py]"]
  node_prompt_builder["Prompt builder<br/>[prompt_builder.py]"]
  node_verifier["Citation verifier<br/>[verifier.py]"]
end

subgraph group_learning["Edit learning"]
  node_learning_route["Edit and preference API<br/>[router.py]"]
  node_edit_capture["Edit capture<br/>[capture.py]"]
  node_pattern_extractor["Preference extraction"]
  node_preference_store[("Preference store")]
end

node_operator(("Legal operator"))
node_backend_app["FastAPI application<br/>[main.py]"]
node_sql_store[("SQLite records<br/>[db.py]")]
node_llm_factory["LLM provider factory<br/>[factory.py]"]
node_groq_provider["Groq provider<br/>[groq_llm.py]"]
node_gemini_provider["Gemini provider<br/>[gemini_llm.py]"]
node_local_embedder["Local embeddings<br/>[local_embedder.py]"]
node_hosted_llm{{"Hosted LLM services"}}

node_operator -->|"uploads documents"| node_upload_ui
node_upload_ui -->|"sends requests"| node_api_client
node_document_view -.->|"loads documents"| node_api_client
node_draft_panel -.->|"requests drafts"| node_api_client
node_draft_panel -->|"renders drafts"| node_draft_renderer
node_draft_panel -->|"compares edits"| node_diff_editor
node_draft_panel -->|"shows preferences"| node_preferences_ui
node_api_client -->|"calls over HTTP"| node_backend_app
node_backend_app -->|"registers routes"| node_ingest_route
node_backend_app -->|"registers routes"| node_rag_route
node_backend_app -->|"registers routes"| node_draft_route
node_backend_app -->|"registers routes"| node_learning_route
node_backend_app -->|"initializes schema"| node_sql_store
node_ingest_route -->|"extracts text"| node_text_extraction
node_text_extraction -->|"uses OCR"| node_ocr
node_ingest_route -->|"extracts fields"| node_structured_extraction
node_ingest_route -->|"chunks pages"| node_chunking
node_ingest_route -->|"indexes chunks"| node_rag_store
node_ingest_route -->|"persists documents"| node_sql_store
node_structured_extraction -->|"uses LLM"| node_llm_factory
node_rag_route -->|"queries index"| node_rag_store
node_drafter -->|"retrieves sources"| node_hybrid_retrieval
node_hybrid_retrieval -->|"searches chunks"| node_rag_store
node_rag_store -->|"embeds text"| node_local_embedder
node_draft_route -->|"generates draft"| node_drafter
node_drafter -->|"builds prompt"| node_prompt_builder
node_drafter -->|"checks grounding"| node_verifier
node_drafter -->|"loads preferences"| node_preference_store
node_drafter -->|"stores draft"| node_sql_store
node_drafter -->|"generates with LLM"| node_llm_factory
node_verifier -->|"checks with LLM"| node_llm_factory
node_learning_route -->|"stores edit"| node_edit_capture
node_learning_route -->|"extracts patterns"| node_pattern_extractor
node_learning_route -->|"saves preferences"| node_preference_store
node_edit_capture -->|"persists edits"| node_sql_store
node_pattern_extractor -->|"analyzes edits"| node_llm_factory
node_preference_store -->|"reads and writes"| node_sql_store
node_llm_factory -->|"selects provider"| node_groq_provider
node_llm_factory -->|"selects provider"| node_gemini_provider
node_groq_provider -->|"calls Groq"| node_hosted_llm
node_gemini_provider -->|"calls Gemini"| node_hosted_llm
node_groq_provider -->|"delegates embeddings"| node_local_embedder
node_gemini_provider -->|"delegates embeddings"| node_local_embedder

click node_upload_ui "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/app/page.tsx"
click node_document_view "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/app/documents/%5Bid%5D/page.tsx"
click node_draft_panel "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DraftPanel.tsx"
click node_draft_renderer "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DraftRenderer.tsx"
click node_diff_editor "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DiffEditor.tsx"
click node_preferences_ui "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/PreferenceList.tsx"
click node_api_client "https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/lib/api.ts"
click node_backend_app "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/main.py"
click node_ingest_route "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/router.py"
click node_text_extraction "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/extractor.py"
click node_ocr "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/ocr.py"
click node_structured_extraction "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/structured.py"
click node_chunking "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/chunker.py"
click node_rag_store "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/store.py"
click node_hybrid_retrieval "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/hybrid.py"
click node_rag_route "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/router.py"
click node_draft_route "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/router.py"
click node_drafter "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/drafter.py"
click node_prompt_builder "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/prompt_builder.py"
click node_verifier "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/verifier.py"
click node_learning_route "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/router.py"
click node_edit_capture "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/capture.py"
click node_pattern_extractor "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/pattern_extractor.py"
click node_preference_store "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/preference_store.py"
click node_sql_store "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/db.py"
click node_llm_factory "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/factory.py"
click node_groq_provider "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/groq_llm.py"
click node_gemini_provider "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/gemini_llm.py"
click node_local_embedder "https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/local_embedder.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_upload_ui,node_document_view,node_draft_panel,node_draft_renderer,node_diff_editor,node_preferences_ui,node_api_client toneBlue
class node_ingest_route,node_text_extraction,node_ocr,node_structured_extraction,node_sql_store toneAmber
class node_chunking,node_rag_store,node_hybrid_retrieval,node_rag_route,node_backend_app toneMint
class node_draft_route,node_drafter,node_prompt_builder,node_verifier toneRose
class node_learning_route,node_edit_capture,node_pattern_extractor,node_preference_store,node_operator,node_llm_factory,node_groq_provider,node_gemini_provider,node_hosted_llm toneIndigo
class node_local_embedder toneTeal
```

## Components

### Review workspace

- [Upload and review](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/app/page.tsx): `frontend/app/page.tsx`
- [Document view](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/app/documents/%5Bid%5D/page.tsx): `frontend/app/documents/[id]/page.tsx`
- [Draft workspace](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DraftPanel.tsx): `frontend/components/DraftPanel.tsx`
- [Citation renderer](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DraftRenderer.tsx): `frontend/components/DraftRenderer.tsx`
- [Edit comparison](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/DiffEditor.tsx): `frontend/components/DiffEditor.tsx`
- [Preference viewer](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/components/PreferenceList.tsx): `frontend/components/PreferenceList.tsx`
- [HTTP client](https://github.com/achemnomaer/legal-draft-ai/blob/main/frontend/lib/api.ts): `frontend/lib/api.ts`

### Document ingestion

- [Ingest and documents](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/router.py): `backend/ingestion/router.py`
- [PDF and OCR extraction](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/extractor.py): `backend/ingestion/extractor.py`
- [OCR processing](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/ocr.py): `backend/ingestion/ocr.py`
- [Structured fields](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/ingestion/structured.py): `backend/ingestion/structured.py`

### Search and grounding

- [Document chunking](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/chunker.py): `backend/rag/chunker.py`
- [Vector index](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/store.py): `backend/rag/store.py`
- [Hybrid retrieval](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/hybrid.py): `backend/rag/hybrid.py`
- [Query endpoint](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/rag/router.py): `backend/rag/router.py`

### Draft generation

- [Draft endpoint](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/router.py): `backend/generation/router.py`
- [Grounded drafter](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/drafter.py): `backend/generation/drafter.py`
- [Prompt builder](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/prompt_builder.py): `backend/generation/prompt_builder.py`
- [Citation verifier](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/generation/verifier.py): `backend/generation/verifier.py`

### Edit learning

- [Edit and preference API](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/router.py): `backend/learning/router.py`
- [Edit capture](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/capture.py): `backend/learning/capture.py`
- [Preference extraction](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/pattern_extractor.py): `backend/learning/pattern_extractor.py`
- [Preference store](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/learning/preference_store.py): `backend/learning/preference_store.py`

### Other components

- **Legal operator**
- [FastAPI application](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/main.py): `backend/main.py`
- [SQLite records](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/db.py): `backend/db.py`
- [LLM provider factory](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/factory.py): `backend/llm/factory.py`
- [Groq provider](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/groq_llm.py): `backend/llm/providers/groq_llm.py`
- [Gemini provider](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/gemini_llm.py): `backend/llm/providers/gemini_llm.py`
- [Local embeddings](https://github.com/achemnomaer/legal-draft-ai/blob/main/backend/llm/providers/local_embedder.py): `backend/llm/providers/local_embedder.py`
- **Hosted LLM services**

## Connections

- Legal operator → Upload and review: uploads documents
- Upload and review → HTTP client: sends requests
- Document view → HTTP client: loads documents
- Draft workspace → HTTP client: requests drafts
- Draft workspace → Citation renderer: renders drafts (evidence: `frontend/components/DraftPanel.tsx`)
- Draft workspace → Edit comparison: compares edits (evidence: `frontend/components/DraftPanel.tsx`)
- Draft workspace → Preference viewer: shows preferences (evidence: `frontend/components/DraftPanel.tsx`)
- HTTP client → FastAPI application: calls over HTTP
- FastAPI application → Ingest and documents: registers routes (evidence: `backend/main.py`)
- FastAPI application → Query endpoint: registers routes (evidence: `backend/main.py`)
- FastAPI application → Draft endpoint: registers routes (evidence: `backend/main.py`)
- FastAPI application → Edit and preference API: registers routes (evidence: `backend/main.py`)
- FastAPI application → SQLite records: initializes schema (evidence: `backend/main.py`)
- Ingest and documents → PDF and OCR extraction: extracts text (evidence: `backend/ingestion/router.py`)
- PDF and OCR extraction → OCR processing: uses OCR (evidence: `backend/ingestion/extractor.py`)
- Ingest and documents → Structured fields: extracts fields (evidence: `backend/ingestion/router.py`)
- Ingest and documents → Document chunking: chunks pages (evidence: `backend/ingestion/router.py`)
- Ingest and documents → Vector index: indexes chunks (evidence: `backend/ingestion/router.py`)
- Ingest and documents → SQLite records: persists documents (evidence: `backend/ingestion/router.py`)
- Structured fields → LLM provider factory: uses LLM (evidence: `backend/ingestion/structured.py`)
- Query endpoint → Vector index: queries index (evidence: `backend/rag/router.py`)
- Grounded drafter → Hybrid retrieval: retrieves sources (evidence: `backend/generation/drafter.py`)
- Hybrid retrieval → Vector index: searches chunks (evidence: `backend/rag/hybrid.py`)
- Vector index → Local embeddings: embeds text (evidence: `backend/rag/store.py`)
- Draft endpoint → Grounded drafter: generates draft (evidence: `backend/generation/router.py`)
- Grounded drafter → Prompt builder: builds prompt (evidence: `backend/generation/drafter.py`)
- Grounded drafter → Citation verifier: checks grounding (evidence: `backend/generation/drafter.py`)
- Grounded drafter → Preference store: loads preferences (evidence: `backend/generation/drafter.py`)
- Grounded drafter → SQLite records: stores draft (evidence: `backend/generation/drafter.py`)
- Grounded drafter → LLM provider factory: generates with LLM (evidence: `backend/generation/drafter.py`)
- Citation verifier → LLM provider factory: checks with LLM (evidence: `backend/generation/verifier.py`)
- Edit and preference API → Edit capture: stores edit (evidence: `backend/learning/router.py`)
- Edit and preference API → Preference extraction: extracts patterns (evidence: `backend/learning/router.py`)
- Edit and preference API → Preference store: saves preferences (evidence: `backend/learning/router.py`)
- Edit capture → SQLite records: persists edits (evidence: `backend/learning/capture.py`)
- Preference extraction → LLM provider factory: analyzes edits (evidence: `backend/learning/pattern_extractor.py`)
- Preference store → SQLite records: reads and writes (evidence: `backend/learning/preference_store.py`)
- LLM provider factory → Groq provider: selects provider (evidence: `backend/llm/factory.py`)
- LLM provider factory → Gemini provider: selects provider (evidence: `backend/llm/factory.py`)
- Groq provider → Hosted LLM services: calls Groq (evidence: `backend/llm/providers/groq_llm.py`)
- Gemini provider → Hosted LLM services: calls Gemini (evidence: `backend/llm/providers/gemini_llm.py`)
- Groq provider → Local embeddings: delegates embeddings (evidence: `backend/llm/providers/groq_llm.py`)
- Gemini provider → Local embeddings: delegates embeddings (evidence: `backend/llm/providers/gemini_llm.py`)
