docs: add document processing recipe
This commit is contained in:
@@ -0,0 +1,36 @@
|
||||
# Document Processing Agent
|
||||
|
||||
## Goal
|
||||
|
||||
Extract structured information (name, date, amount) from unstructured text or documents.
|
||||
|
||||
## Nodes
|
||||
|
||||
### 1. Input Node
|
||||
|
||||
- Accept raw text or document content
|
||||
|
||||
### 2. Extraction Node
|
||||
|
||||
- Use LLM or parsing logic to extract:
|
||||
- name
|
||||
- date
|
||||
- amount
|
||||
|
||||
### 3. Output Node
|
||||
|
||||
- Return structured JSON
|
||||
|
||||
## Edges
|
||||
|
||||
- Input → Extraction → Output
|
||||
|
||||
## Tools
|
||||
|
||||
- LLM (OpenAI / Anthropic)
|
||||
- Optional: OCR for PDFs
|
||||
|
||||
## Usage notes
|
||||
|
||||
- Useful for invoice processing
|
||||
- Can be extended for contracts, forms, etc.
|
||||
Reference in New Issue
Block a user