# Contract Audit

**Route:** `/audit` **File:** `app/audit/page.tsx`

The Audit page is the starting point for every contract workflow. It analyzes a contract for fairness, risks, and price accuracy using local QVAC AI — in under 30 seconds.

***

## Input Methods

### 1. Upload PDF

* Drag-and-drop or click to upload a `.pdf` file
* Sent to `/api/upload`, extracted via `pdf-parse`, raw text returned
* Best for: contracts received as official documents

### 2. Paste Text

* Click the **Text** tab and paste raw contract text directly
* Best for: contracts in Google Docs, email, or any digital format

### 3. Demo Contract

* Click **Coba Demo Kontrak** to load a pre-loaded sample contract (`D:\frontier\RAB_Pengembangan_Website.pdf`) via `/api/demo-pdf`
* Best for: quick testing without a real contract file

***

## Analysis Process

```mermaid
flowchart LR
    A([Contract Text]) --> B([Detect\nContract Type])
    B --> C([Fetch Market\nPrices])
    B --> D([QVAC Analyzes\nwith Expert Persona])
    C --> D
    D --> E([Structured\nJSON Result])
    E --> F([Rendered\nin UI])
```

**Time:** 3–20 seconds depending on QVAC model tier and contract length.

Analysis runs through `/api/audit-stream`, which streams live status updates via SSE so the user sees progress instead of a blank screen.

***

## Output Sections

### Fairness Score

A single score from **1–10** representing overall contract balance:

| Score | Rating     | Color  | Recommended Action     |
| ----- | ---------- | ------ | ---------------------- |
| 8–10  | Fair       | Green  | Safe to proceed        |
| 5–7   | Concerns   | Yellow | Review flagged clauses |
| 1–4   | Imbalanced | Red    | Request revisions      |

***

### Risky Clauses

Each flagged clause includes:

* **Risk level** badge: `HIGH` / `MEDIUM` / `LOW`
* **Exact clause text** from the contract
* **Plain-language explanation** of why it's risky
* **Suggested revision** to make it fair

***

### Price Analysis

Line-by-line comparison against real Indonesian market data:

* Item name and contracted price
* Market price range (fetched from Blibli via FastAPI backend, and optionally Google Shopping / SerpAPI)
* Status badge: `OVERPRICED` / `FAIR` / `UNDERPRICED`
* Notes explaining the discrepancy from both perspectives (client and contractor)

***

### Regulation Compliance

Checks relevant Indonesian laws based on the detected contract type. Each check shows:

* Regulation reference (e.g., *Perpres No. 16/2018 Pasal 7*)
* Compliance status: Compliant / Non-compliant
* Specific notes on what needs to change

***

### Revision Suggestions & Uncertainty Questions

* **Revision Suggestions** — actionable bullet points for improving contract balance
* **Uncertainty Questions** — ambiguous clauses the AI flagged for clarification with the other party

***

## Audit Hash

After analysis, a **SHA-256 hash** of the audit result is displayed.

This hash:

* Proves the contract was audited before deployment
* Is recorded on-chain when you deploy the contract
* Can be used to verify audit integrity at any time

***

## Contract Chat (Q\&A)

After auditing, a **Chat** tab appears on the page. You can ask any question about the contract in plain language:

> *"Does this contract have a late payment penalty clause?"* *"What are my IP rights under Clause 5?"*

Powered by `/api/chat-contract`. The prior audit result is automatically passed as context for better answers.

***

## Next Step

After auditing, deploy the contract → [Create Contract](/contractguard-ai-docs/features/create-contract.md)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://contractguard-ai.gitbook.io/contractguard-ai-docs/features/contract-audit.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.