wxlawyers/lawyer-knowledge-graph
Developer toolsThis page was translated automatically. Technical details may differ from the original. View the original
34 professional skills · Covering the entire case handling process · Boosting lawyer efficiency by 10 times
Overview
An AI personal knowledge management system for litigation lawyers + 40 professional skill sets. Two modules complement each other: - : Automatic accumulation of case experience → vector retrieval → second-level matching of case types - : Hermes Agent skills covering the entire case handling process, ready for conversation A few years ago, I encountered a sales contract dispute. The facts were almost identical to another case I handled five years ago—the same focus of contention, the same legal application issue. But after searching through folders, I couldn't find any detailed records from that time, only remembering that "it seemed like something similar had been handled," and finally restarted the study, spending three extra days. Lawyers' case-handling experience is being wasted. The understanding accumulated with each case—how to identify the key points of dispute, how to organize the chain of evidence, how the court takes on similar issues—these experiences are either scattered in Word documents or sleeping in memory. Each knowledge card automatically generates a vector index and stores it in the pgvector database. The more cases entered, the richer the system's accumulated knowledge. When receiving a new case, enter a case description in the search box, and the system will search all relevant knowledge cards from historical cases based on ** (rather than keyword matching): Set focus areas (such as "contract disputes" and "labor disputes"), automatically search for newly added judicial interpretations and guiding cases, and generate updated summaries for push notifications. Reserved interfaces: PKULAW, YUANDIAN, which can be enabled by configuring the API Key.
Overview (machine translated)
**Lawyer Knowledge Graph — Turning Every Case Handled into an Asset for Future Case Handling **
An AI personal knowledge management system for litigation lawyers + 40 professional skill sets. Two modules complement each other:
- Knowledge Graph System: Automatic case experience accumulation → vector retrieval→ second-level case matching
- Skill Set: Hermes Agent skills covering the entire case-handling process, ready to use in conversation
A pain point
Having worked as a civil and commercial litigation lawyer for fifteen years, he has handled hundreds of cases.
A few years ago, I encountered a sales contract dispute. The facts were almost identical to another case I handled five years ago—the same focus of contention, the same legal application issue. But after searching through folders, I couldn’t find any detailed records from that time, only remembering that “it seemed like something similar had been handled,” and finally restarted the study, spending three extra days.
Lawyers’ case-handling experience is being wasted. The understanding accumulated with each case—how to identify the key points of dispute, how to organize the chain of evidence, how the court takes on similar issues—these experiences are either scattered in Word documents or sleeping in memory.
**This is where the Lawyer Knowledge Graph begins. **
Project Architecture
Module One: Knowledge Graph System
Core Logic
Three Core Features
1. Case knowledge is automatically accumulated
Paste judgment documents or case file summaries, and AI automatically extracts structured knowledge cards:
| Card Type | Description | Example |
|---|---|---|
| Focus of Dispute | Core Disputed Issues in This Case | “Should the Liquidated Damages Penalty Be Reduced If Too High?” |
| Judicial Rules | Court Standards for Adjudicating Such Issues | “Liquidated damages reduced to 30% of actual losses as a reference cap” |
| Application of Legal Provisions | Precise to the specific law number and key points | “Article 585 of the Civil Code: Rules for Adjusting Liquidated Damages” |
| Case Handling Experience | Practical Skills from a Lawyer’s Perspective | “Such cases should focus on collecting actual loss certificates and industry profit rate data” |
Each knowledge card automatically generates a vector index and stores it in the pgvector database. The more cases entered, the richer the system’s accumulated knowledge.
2. Intelligent case search
When receiving a new case, enter the case description in the search box. The system retrieves all relevant knowledge cards from historical cases based on semantic similarity (rather than keyword matching):
3. Regulatory Dynamic Tracking (in Planning)
Set focus areas (such as “contract disputes” and “labor disputes”), automatically search for newly added judicial interpretations and guiding cases, and generate updated summaries for push notifications.
Reserved interfaces: PKULAW, YUANDIAN, which can be enabled by configuring the API Key.
Technical Architecture
| Components | Technical Selection | Notes |
|---|---|---|
| Backend Framework | FastAPI | Asynchronous high performance, automatically generating API documentation |
| Database | PostgreSQL 16 + pgvector | Structured data + vector retrieval |
This page was translated automatically. Technical details may differ from the original.
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add wxlawyers/lawyer-knowledge-graph