---
generator: "vaultspec-marketing discovery"
title: "When search results look wrong - vaultspec-rag documentation"
description: "Search reads the stored index. Start with an installed and provisioned workspace."
url: "https://vaultspec.neve.md/docs/rag/verification.html"
component: "vaultspec-rag"
source_repository: "https://github.com/nevenincs/vaultspec-rag"
source_path: "docs/verification.md"
source_commit: "9093ad50f015110cf0061432da49ce859282c7d6"
date_modified: "2026-09-30T14:43:23+02:00"

---

[View as Markdown](<https://vaultspec.neve.md/docs/rag/verification.md>)

# When search results look wrong

Search reads the stored index. Start with an [installed and provisioned workspace](<https://vaultspec.neve.md/docs/rag/installation.html>).

Examples use the installed-tool form and call `vaultspec-rag` directly. If vaultspec-rag is a project dependency, prefix each command with `uv run`; see the [installation guide](<https://vaultspec.neve.md/docs/rag/installation.html>) for lane selection, including the standalone tool and no-install routes.

## Is the service healthy?

**Command**

```bash
vaultspec-rag server doctor
```

Inspect `Live service`, `Installed dependencies`, and, when present, `Provisioning (vaultspec-rag)`.

- For dependency faults, follow the [installation troubleshooting guide](<https://vaultspec.neve.md/docs/rag/installation.html#when-something-goes-wrong>).
- For service faults, follow the [service troubleshooting guide](<https://vaultspec.neve.md/docs/rag/service-mode.html#troubleshooting>).
- For provisioning faults, read the reported detail.

If none of these checks reports a problem, [check index status](<https://vaultspec.neve.md/docs/rag/verification.html#check-index-status>).

## Check index status

**Command**

```bash
vaultspec-rag status --verbose
```

Counts describe stored sections, not source files, and don’t prove the index is current.

Add `--verbose` to see the `Index generations` section. Under it, each domain shows its latest recorded indexing job for this project. `not indexed yet` means no matching job was found in the retained history; it doesn’t establish that the store is empty. `succeeded` records that job’s success, not whether the index matches your current files.

To index the domains you need, follow [reindexing](<https://vaultspec.neve.md/docs/rag/verification.html#reindexing>). If expected files are missing, [check file coverage](<https://vaultspec.neve.md/docs/rag/verification.html#is-it-indexing-the-right-files>).

## Is it indexing the right files?

Preview which source files would be indexed without changing the index:

**Command**

```bash
vaultspec-rag index --type code --dry-run
```

Check that expected files appear and unwanted files stay out. The admission summary counts admitted and rejected files by reason.

By default, the preview displays up to 50 paths. Use `--dry-run-limit N` to change that limit, or add `--json` for the full admitted path list.

If file selection is wrong, review the [admission rules](<https://vaultspec.neve.md/docs/rag/indexing.html#indexing-pipeline>) and [routing configuration](<https://vaultspec.neve.md/docs/rag/preprocessing-hooks.html#configure-rules>). Repeat the preview after making changes.

## Reindexing

To update the index after changing files or indexing settings:

**Command**

```bash
vaultspec-rag index
```

Scope it with `--type vault`, `--type code`, or `--type document` when only one kind has changed. To delete and rebuild a domain, pass both `--type` and `--rebuild`.

The running service reindexes as files change. See [watcher operation](<https://vaultspec.neve.md/docs/rag/automation.html#automatic-re-indexing>) for automatic updates.

## Gating a script on these checks

See [scripting and automation](<https://vaultspec.neve.md/docs/rag/automation.html>) for JSON output and exit-code handling.

## Related documentation

- [Storage maintenance](<https://vaultspec.neve.md/docs/rag/storage-maintenance.html>) covers pruning what indexing leaves behind.
- [Search and index](<https://vaultspec.neve.md/docs/rag/search-and-index.html>) covers the search and index commands in full.
