# VantaDB Documentation Registry
Welcome to the VantaDB documentation registry. This index maps all available guides, policies, design records, and technical specifications, classified by target audience and domain, adhering to a strict 5-pillar structure.
---
## 🚀 Getting Started & Releases
* [QUICKSTART.md](QUICKSTART.md) — 5-minute setup guide for Rust, CLI client, and Python SDK.
* [Changelog](CHANGELOG.md) — Summary of releases and stable versions history.
---
## 🚚 Migration Guides
* [From ChromaDB](migration/FROM_CHROMADB.md) — Step-by-step guide to migrate from ChromaDB to VantaDB.
* [From LanceDB](migration/FROM_LANCEDB.md) — Step-by-step guide to migrate from LanceDB to VantaDB.
---
## 🧠 GraphRAG (Knowledge Graph + Retrieval)
* [GraphRAG on VantaDB](graphrag/README.md) — Unified vector-graph database for GraphRAG: semantic clusters, confidence scoring, graph traversal, semantic compression engine, and multi-agent support.
---
## 💻 API & SDK Reference (Pillar 1)
* [Python SDK Guide](api/PYTHON_SDK.md) — Integration, GIL release policies, Rayon batching, and PyO3 FFI boundary specs.
* [Model Context Protocol (MCP)](api/MCP.md) — Guide to using the Model Context Protocol (MCP) server for connecting local AI agents directly.
---
## 🏛️ Architecture & System Design (Pillar 2)
* [Architecture Overview](architecture/ARCHITECTURE.md) — Core architecture principles (Single-crate design, layout alignment, and zero-copy memory mapping).
* [Text Index Design](architecture/TEXT_INDEX_DESIGN.md) — System design of the inverted text index for BM25 lexical search.
* [WAL & Mutation Recovery](architecture/MUTATION_RECOVERY_PROTOCOL.md) — WAL specifications, CRC32C validation, and the Scan-Forward Auto-healing protocol.
* [Advanced Tokenizer](architecture/ADVANCED_TOKENIZER.md) — Multilingual text processing, stemming, and Unicode folding based on Tantivy.
* [Architecture Decision Records (ADRs)](architecture/adr/) — Formal record of architectural decisions:
* [ADR 001: Unified Config and Read-Only Mode](architecture/adr/001_unified_config_readonly.md)
* [ADR 002: WAL CRC32C and Scan-Forward Auto-healing](architecture/adr/002_wal_crc32c_autohealing.md)
* [ADR 003: Sync-Async Core Decoupling](architecture/adr/003_sync_async_decoupling.md)
* [Historical Audits](architecture/audits/) — Technical review logs from early development phases:
* [2026-05-04 Cleanup Candidates](architecture/audits/2026-05-04-cleanup-candidates.md)
* [2026-05-04 Test Report](architecture/audits/2026-05-04-test-report.md)
* [2026-05-04 Total Review](architecture/audits/2026-05-04-total-review.md)
* [2026-05-19 Phase 5 Certification Report](architecture/audits/2026-05-19-fase-5-certification-report.md)
* [2026-05-19 Performance Action Plan](architecture/audits/2026-05-19-plan-accion-alto-rendimiento.md)
---
## ⚙️ Operations, Policies & Performance (Pillar 3)
* [Executive Technical Audit](operations/EXECUTIVE_TECHNICAL_AUDIT.md) — Executive summary of architectural review, findings, and remediation (Unified technical audit report).
* [Benchmarks & Performance](operations/BENCHMARKS.md) — Performance metrics, memory footprints, and competitive analysis vs. LanceDB and ChromaDB.
* [Configuration Schema](operations/CONFIGURATION.md) — Detail of parameters in `VantaConfig`, cache limits, and query boundaries.
* [Memory Telemetry](operations/MEMORY_TELEMETRY.md) — Metrics for tracking physical RAM RSS and mapped page tables (Unified telemetry baseline).
* [Backup & Durability Policy](operations/BACKUP_POLICY.md) — Hot checkpoints, filesystem backups, and index snapshot strategies.
* [Continuous Integration (CI)](operations/CI_POLICY.md) — Rules for PR testing gates, coverage constraints, and lint enforcement.
* [Python Release Policy](operations/PYTHON_RELEASE_POLICY.md) — Automated wheels compilation via cibuildwheel and signing with GitHub Attestations (SLSA Level 2).
* [Product & Feature Boundary](operations/EXPERIMENTAL_FEATURES.md) — Categorized matrix of stable (MVP), optional wrappers, and experimental subsystems.
* [Reliability & Chaos Testing](operations/RELIABILITY_GATE.md) — Test suites for chaos loop, failpoints injection, and hardware profiles.
* [Fuzzing Guide](operations/FUZZING.md) — Harnesses and guidelines for running fuzz tests using cargo fuzz.
* [Repository Checklist](operations/REPO_CHECKLIST.md) — Code hygiene, documentation and packaging check sheets.
* [Editor Setup Guidelines](operations/EDITOR_INTEGRATIONS.md) — Editor setup guidelines for VS Code and NeoVim.
* [Agent Guidelines](operations/AGENT_INSTRUCTIONS.md) — Directives and rules of governance for AI coding agents operating in this repository.
* [Unified Progress History](progreso/README.md) — Consolidated summary of all development phases and milestones completed.
* [Historical Repository Snapshots](operations/snapshots/) — Full backups of codebase status at key points:
* [Snapshot 2026-06-09](operations/snapshots/snapshot_2026-06-09.md)
* [Snapshot 2026-06-12](operations/snapshots/snapshot_2026-06-12.md)
---
## 👥 Case Studies & Community (Pillar 4)
* [Case Studies](case_studies/) — Real-world deployment scenarios:
* [Local Memory with Ollama](case_studies/agent_local_memory_ollama.md) — Embedding integration with local LLMs.
* [RAG on Edge Devices](case_studies/rag_edge_device.md) — Edge deployments on resource-restricted systems.
* [Pilot Program Plan](operations/PILOT_PROGRAM.md) — Framework and guidelines for early developer pilots (Unified pilot pack & outreach plan).
* [Community Governance](operations/COMMUNITY_GOVERNANCE.md) — Contribution process, RFC timelines, and Code of Conduct.
* [HackerNews Launch Preparation](operations/SHOW_HN_PREP.md) — Post copywriting drafts and Q&A to answer technical critiques.
* [Public Issues & Good First Issues](operations/PUBLIC_ISSUE_DRAFTS.md) — Automated templates for first-time open-source contributors.
---
## 🔬 Experimental (Pillar 5)
* [Experimental IQL](experimental/IQL.md) — Specifications of the experimental LISP-like query parser currently in quarantine.
---
## ✍️ Technical Articles & Publications
* [How Hybrid Search Works](articles/how_hybrid_search_works.md) — Deep dive into combining Tantivy lexical search with HNSW vector search.
* [SQLite for AI Agents](articles/sqlite_for_ai_agents.md) — Why a specialized vector-relational engine is needed instead of traditional SQLite.
* [Why I Built a Local Memory Engine](articles/why_i_built_local_memory_engine.md) — Engineering rationale behind creating a zero-dependency local vector database.
---
## 📂 Active Checklist & Branch Progress
* [Unified Progress History](progreso/README.md) — Consolidated log of completed phases. Historical subdirectories are deleted after consolidation to keep the repository clean.