⚠️ Early Development — API may change.

What is Vectorless?

Vectorless is a Rust library for querying structured documents using natural language — without vector databases or embedding models.

Instead of chunking documents into vectors, Vectorless preserves the document's tree structure and uses a hybrid algorithm + LLM approach to navigate it — like how a human reads a table of contents:

• Algorithm handles "how to walk" — BM25 scoring, tree traversal (fast, deterministic)
• Pilot (LLM) handles "where to go" — semantic understanding, ambiguity resolution

Analogy: Traditional RAG is like searching every word in a book. Vectorless is like reading the table of contents, then going to the right chapter.

How It Works

1. Index: Build a Navigable Tree

Technical Manual (root)
├── Chapter 1: Introduction
├── Chapter 2: Architecture
│   ├── 2.1 System Design
│   └── 2.2 Implementation
└── Chapter 3: API Reference

Each node gets an AI-generated summary, enabling fast navigation.

2. Query: Navigate with LLM

When you ask "How do I reset the device?":

1. Analyze — Understand query intent and complexity
2. Navigate — LLM guides tree traversal (like reading a TOC)
3. Retrieve — Return the exact section with context
4. Verify — Check if more information is needed (backtracking)

Traditional RAG vs Vectorless

Aspect	Traditional RAG	Vectorless
Infrastructure	Vector DB + Embedding Model	Just LLM API
Document Structure	Lost in chunking	Preserved
Context	Fragment only	Section + surrounding context
Setup Time	Hours to Days	Minutes
Best For	Unstructured text	Structured documents

Example

Input:

Document: 100-page technical manual (PDF)
Query: "How do I reset the device?"

Output:

Answer: "To reset the device, hold the power button for 10 seconds 
until the LED flashes blue, then release..."

Source: Chapter 4 > Section 4.2 > Reset Procedure

When to Use

✅ Good fit:

• Technical documentation
• Manuals and guides
• Structured reports
• Policy documents
• Any document with clear hierarchy

❌ Not ideal:

• Unstructured text (tweets, chat logs)
• Very short documents (< 1 page)
• Pure Q&A datasets without structure

Quick Start

Installation

[dependencies]
vectorless = "0.1"

Configuration

cp vectorless.example.toml ./vectorless.toml

Usage

use vectorless::Engine;

#[tokio::main]
async fn main() -> vectorless::Result<()> {
    // Create client
    let client = Engine::builder()
        .with_workspace("./workspace")
        .build()?;

    // Index a document (PDF, Markdown, DOCX, HTML)
    let doc_id = client.index("./document.pdf").await?;

    // Query with natural language
    let result = client.query(&doc_id, "What are the system requirements?").await?;

    println!("Answer: {}", result.content);
    println!("Source: {}", result.path); // e.g., "Chapter 2 > Section 2.1"

    Ok(())
}

Features

Feature	Description
Zero Infrastructure	No vector DB, no embedding model — just an LLM API
Multi-format Support	PDF, Markdown, DOCX, HTML out of the box
Incremental Updates	Add/remove documents without full re-index
Traceable Results	See the exact navigation path taken
Feedback Learning	Improves from user feedback over time
Multi-turn Queries	Handles complex questions with decomposition

Architecture

Core Components

• Index Pipeline — Parses documents, builds tree, generates summaries
• Retrieval Pipeline — Analyzes query, navigates tree, returns results
• Pilot — LLM-powered navigator that guides retrieval decisions
• Metrics Hub — Unified observability for LLM calls, retrieval, and feedback

Examples

See the examples/ directory.

Contributing

Contributions welcome! If you find this useful, please ⭐ the repo — it helps others discover it.

Star History

License

Apache License 2.0