API Reference

1 API Reference

This reference covers VeritaScribe’s programmatic interfaces, data models, and internal APIs for developers and advanced users.

1.1 Command Line Interface

1.1.1 Core Commands

1.1.1.1 analyze

Perform comprehensive thesis analysis.

uv run python -m veritascribe analyze [OPTIONS] PDF_PATH

Arguments: - PDF_PATH (required): Path to the PDF thesis file

Options: - --output, -o TEXT: Output directory for reports - --citation-style, -c TEXT: Expected citation style (APA, MLA, Chicago, IEEE, Harvard) - --quick, -q: Perform quick analysis (first 10 blocks only) - --no-viz: Skip generating visualization charts - --annotate: Generate an annotated PDF with highlighted errors - --verbose, -v: Enable verbose logging

Exit Codes: - 0: Success - 1: Error (file not found, analysis failed, etc.)

1.1.1.2 quick

Fast analysis of document subset.

uv run python -m veritascribe quick [OPTIONS] PDF_PATH

Arguments: - PDF_PATH (required): Path to the PDF thesis file

Options: - --blocks, -b INTEGER: Number of text blocks to analyze (default: 5)

1.1.1.3 demo

Create and analyze sample document.

uv run python -m veritascribe demo

No arguments or options. Creates demo_thesis.pdf in current directory.

1.1.1.4 config

Display current configuration.

uv run python -m veritascribe config

Outputs configuration table and API key status.

1.1.1.5 providers

List available LLM providers and models.

uv run python -m veritascribe providers

1.1.1.6 optimize-prompts

Fine-tune analysis prompts using a few-shot learning dataset.

uv run python -m veritascribe optimize-prompts

1.1.1.7 test

Run system diagnostics.

uv run python -m veritascribe test

Tests configuration loading, PDF processing, and LLM connectivity.

1.2 Python API

1.2.1 Configuration Management

1.2.1.1 veritascribe.config.VeritaScribeSettings

Main configuration class using Pydantic Settings.

from veritascribe.config import VeritaScribeSettings

settings = VeritaScribeSettings(
    openai_api_key="your-key-here",
    default_model="gpt-4",
    max_tokens=2000,
    temperature=0.1
)

Key Fields:

Field Type Default Description
llm_provider str "openai" LLM provider (openai, openrouter, anthropic, custom)
openai_api_key Optional[str] None OpenAI API key
openrouter_api_key Optional[str] None OpenRouter API key
anthropic_api_key Optional[str] None Anthropic API key
default_model str "gpt-4" LLM model name
max_tokens int 2000 Max tokens per request
temperature float 0.1 LLM temperature

1.2.2 Data Models

All data models use Pydantic for validation and serialization.

1.2.2.1 veritascribe.data_models.ThesisAnalysisReport

Complete analysis report containing all results.

Key Fields: - document_name: str - analysis_timestamp: datetime - total_pages: int - total_words: int - total_text_blocks: int - analysis_results: List[AnalysisResult] - total_processing_time_seconds: float - token_usage: Optional[Dict[str, int]] - estimated_cost: Optional[float]

This API reference covers the main interfaces. For implementation details, see the Architecture Guide.