# Vaultspec published pages > Markdown generated from the published HTML. Each imported page names its source in its metadata. Every documentation page also has a Markdown twin at the same URL stem with a `.md` suffix. [Structured catalog](https://vaultspec.neve.md/content-index.json) ## Documentation home - [Start using vaultspec](https://vaultspec.neve.md/docs/index.md): vaultspec keeps research, decisions, and plans with your project instead of losing them in agent conversations. How vaultspec works explains the records. ## Get started - [Requirements](https://vaultspec.neve.md/docs/requirements.md): Start with Core. It installs the workflow and manages records without vaultspec-rag, a GPU, or an API key. The quickstart follows that path. - [Install vaultspec-core](https://vaultspec.neve.md/docs/core/installation.md): For the full setup path and a working example, start at the documentation home. This page covers installation choices and updates. - [Install vaultspec-rag](https://vaultspec.neve.md/docs/rag/installation.md): vaultspec-rag adds search by meaning over project records and code. - [Your first search](https://vaultspec.neve.md/docs/rag/getting-started.md): Index your project, search its source code, and narrow the results to a file. ## Guides - [How vaultspec works](https://vaultspec.neve.md/docs/concepts.md): Vaultspec stores the reasoning and verification for a feature beside its code. One feature tag connects the records. - [Run a feature](https://vaultspec.neve.md/docs/guides/workflow.md): Start from the documentation home to install Core and create a record. Then connect your agent so it can read and update project records. - [Use the command line](https://vaultspec.neve.md/docs/guides/cli.md): Run these commands from your project root after installing Core. The examples use uvx; change the prefix if you chose a tool or project installation. - [Connect your agent](https://vaultspec.neve.md/docs/guides/mcp.md): Core exposes project operations through the Model Context Protocol (MCP). Your coding agent calls these tools to find records and report progress. - [How vaultspec-rag runs](https://vaultspec.neve.md/docs/guides/rag-service.md): vaultspec-rag is two programs you set up separately. You install a search service and start it yourself. - [TypeSafe and environment loading](https://vaultspec.neve.md/docs/guides/typesafe-env.md): TypeSafe is optional. Core and RAG each read their own API key for their hosted features. Installing Core and managing records needs neither. ## vaultspec-core - [vaultspec-core](https://vaultspec.neve.md/docs/core/index.md): Core installs the workflow, configures supported coding agents, and manages the feature records under .vault/. - [The workflow in depth](https://vaultspec.neve.md/docs/core/framework.md): Start from a configured project, then describe your task to the coding agent. You don’t need to create a complete set of documents before work begins. - [Document syntax](https://vaultspec.neve.md/docs/core/syntax.md): Create documents with vaultspec-core vault add; record execution with vaultspec-core vault exec log, which creates the plan’s ledger when needed. - [Review and correctness](https://vaultspec.neve.md/docs/core/correctness.md): Review the implementation against its approved scope, any governing architecture decision records (ADRs), and test evidence. - [Check a workspace](https://vaultspec.neve.md/docs/core/verification.md): These checks inspect installation and record consistency. - [MCP server](https://vaultspec.neve.md/docs/core/mcp.md): The server exposes Vaultspec document and workflow tools to Model Context Protocol (MCP) clients using JSON-RPC over standard input and output. - [Distribution channels](https://vaultspec.neve.md/docs/core/channels.md): Install Core with Scoop on Windows or Homebrew on macOS and Linux. Both channels install vaultspec-core and vaultspec-core-mcp. ## vaultspec-rag - [vaultspec-rag](https://vaultspec.neve.md/docs/rag/index.md): vaultspec-rag indexes vaultspec records, source code, and routed documents for search by meaning. ### Search - [Search and index](https://vaultspec.neve.md/docs/rag/search-and-index.md): vaultspec-rag searches vault records, source code, and explicitly routed extracted documents by meaning. - [Write a query that finds it](https://vaultspec.neve.md/docs/rag/query-craft.md): Describe the behavior you want to find, then narrow the search with filters. - [When results look wrong](https://vaultspec.neve.md/docs/rag/verification.md): Search reads the stored index. Start with an installed and provisioned workspace. ### Run the service - [The background service](https://vaultspec.neve.md/docs/rag/service-mode.md): Run vaultspec-rag as a long-lived background service to keep the models loaded and the managed server running. - [Models and index-server provisioning](https://vaultspec.neve.md/docs/rag/provisioning.md): A host installation needs two things that do not come with the Python package: the model files and, for the default backend, a Qdrant server executable. - [Automatic index updates](https://vaultspec.neve.md/docs/rag/automatic-convergence.md): Automatic convergence keeps the code, vault, and document indexes current after filesystem changes. - [Service discovery](https://vaultspec.neve.md/docs/rag/service-discovery.md): The resident background daemon is a machine singleton: exactly one may run per machine, because it owns the single GPU and the single managed Qdrant… - [Storage backends](https://vaultspec.neve.md/docs/rag/backends.md): vaultspec-rag uses a managed local Qdrant server by default. Choose local-only storage if you cannot run a separate Qdrant process. - [Storage and maintenance](https://vaultspec.neve.md/docs/rag/storage-maintenance.md): vaultspec-rag keeps semantic search indexes so agents and operators can search your projects by meaning instead of by exact keyword. ### Integrate - [MCP clients](https://vaultspec.neve.md/docs/rag/mcp.md): An assistant like Claude Desktop or Claude Code can search your source code, and the decision records in your .vault/ directory, without you switching… - [Scripting and automation](https://vaultspec.neve.md/docs/rag/automation.md): Pass --json to get machine-readable output instead of a formatted table. - [Preprocessing hooks](https://vaultspec.neve.md/docs/rag/preprocessing-hooks.md): vaultspec-rag keeps vault records, conventional source code, and extracted documents in separate index domains. - [Configuration](https://vaultspec.neve.md/docs/rag/configuration.md): This page lists every VAULTSPEC_RAG_* environment variable vaultspec-rag declares. ### Internals - [Indexing and retrieval](https://vaultspec.neve.md/docs/rag/indexing.md): This page explains how vaultspec-rag turns vault documents and source files into a searchable index, and why each part of the retrieval path is shaped… - [Architecture](https://vaultspec.neve.md/docs/rag/architecture.md): vaultspec-rag is the retrieval layer of a vaultspec-core project. ## Reference - [vaultspec-core CLI](https://vaultspec.neve.md/docs/core/cli.md): Complete command-line interface (CLI) reference for vaultspec-core. See the framework manual for workflows and concepts. - [vaultspec-rag CLI](https://vaultspec.neve.md/docs/rag/cli.md): Generated from the live command surface. Each entry lists the command’s arguments, options, types, defaults, and help text.