The short version
Every tool here can open an object store and show you rows. If that is all you need, install whichever one you like the look of and stop reading, they are all free.
The differences show up once the store is large, the data is not plain JSON, or you need to ask a question more specific than “show me everything”. That is where a table viewer and a database client stop being the same category of thing.
IdxBeaver is the only one of the six that ships a query language or infers a schema, and the only one that covers Cookies and Cache Storage alongside IndexedDB. Kahuna is the closest thing to a peer: it also handles the awkward structured-clone types, exports in Dexie’s format, runs on Firefox, and can edit a schema, which IdxBeaver cannot. If you want the honest counter-argument, skip to when not to use IdxBeaver.
Capability matrix
Claims for the other five tools were taken from their public READMEs and Chrome Web Store listings, checked on 10 September 2026. “Not documented” means the capability is absent from those sources, not that it has been tested and found missing. If one of these is wrong or out of date, open an issue and it will be corrected.
| Capability | IdxBeaver | Kahuna | IndexedDB Browser | idb-crud | IndexedDB Explorer | IndexedDBEdit |
|---|---|---|---|---|---|---|
| Storage surfaces | IndexedDB, LocalStorage, SessionStorage, Cookies, Cache Storage | IndexedDB only | IndexedDB only | IndexedDB, LocalStorage, SessionStorage | IndexedDB only | IndexedDB only |
| Where it runs | DevTools panel | Overlay on the page, outside DevTools | DevTools panel | Drawer UI, outside DevTools | Not stated on listing | DevTools panel |
| Query language | MongoDB-style filter, projection, sort, limit | A JavaScript console over Dexie's API; no declarative query | Not documented | Not documented | Not documented | Not documented |
| Index-aware query planning | Picks an IDBIndex when the filter allows, and reports the plan it used | Not documented | Not documented | Not documented | Not documented | Not documented |
| Saved queries and history | Per origin, auto-trimmed to the 100 most recent | Not documented | Not documented | Not documented | Not documented | Not documented |
| Schema inference | Samples rows for field types and coverage; powers autocomplete and a Structure view | Not documented | Not documented | Not documented | Not documented | Not documented |
| Schema export to code | TypeScript interface or Dexie schema | Not documented | Not documented | Not documented | Not documented | Not documented |
| Data export formats | JSON, NDJSON, CSV, SQL INSERT, ZIP snapshot | Dexie, JSON, CSV | Not documented | Export and import, formats unspecified | Not documented | Not documented |
| Non-JSON value round-trip | Date, BigInt, RegExp, Map, Set, ArrayBuffer, Blob, circular refs | Documents editing Dates, Maps, Sets, RegExps, typed arrays, Blobs, Files and ImageData | Documents sets, maps and typed arrays as unsupported | Not documented | Not documented | Not documented |
| Large object stores | Virtualized grid; rows render on demand | Paginated grid, page size not stated | Documents slower loads for stores with hundreds of thousands of objects | Not documented | Paged, up to 5,000 records per page | Not documented |
| Multi-frame discovery | Scans every scriptable frame and merges results | Not documented | Not documented | Not documented | Not documented | Not documented |
| Schema editing | Delete a database or store; no create, add or alter | Create and delete databases, add and remove object stores and indexes | Not documented | Not documented | Not documented | Not documented |
| Browsers | Chromium 120 or newer | Chromium and Firefox | Chrome Web Store only | Chrome Web Store only | Chrome Web Store only | Chrome Web Store only |
| Editing model | Inline grid edits with per-cell undo and redo | Full create, read, update, delete | Add, edit and delete in a table | Full create, read, update, delete | Add or delete records | Modify or delete key-value records |
| Search, sort, column control | Filter bar, sort, column pin and resize, sticky headers | Filters with regex, sortable grid, configurable, reorderable and hideable columns | Table search, per-column filter, drag-to-reorder columns | Advanced sorting and filtering, customizable columns | Real-time record search | Not documented |
| Source available | Yes, MIT | Yes, on GitHub | Yes, on GitHub | Described as open source | Not stated on listing | Yes, on GitHub |
What the differences look like in practice
A table viewer answers “what is in here”
Five of these six tools are built around one interaction: pick a store, render it as a table, click a cell to change it. That is genuinely the right shape for most debugging. You have twelve rows of app settings and one of them is wrong.
It stops working at scale. Once a store holds tens of thousands of records, “render it as a table and search visible columns” means paging through data hunting for the row you want. IndexedDB Browser is upfront about this in its own README, noting it is designed for small to medium stores and that loads get slower in the hundreds of thousands. IndexedDB Explorer takes the other approach and pages at up to 5,000 records at a time.
A query answers “which rows match”
IdxBeaver treats the store as something you interrogate rather than scroll:
The planner looks for single-field equality and range expressions it can push down to an IDBIndex, and falls back to an in-memory match for compound operators. It then tells you which path it took, which is how you notice a store is missing an index it should have. No other tool in this comparison documents anything equivalent.
Types are where naive viewers quietly lose data
IndexedDB stores structured-clone values, not JSON. A record can hold a Date, a Map, a Set, a BigInt, an ArrayBuffer, a Blob, or a circular reference. Render that through JSON.stringify and a Date becomes a string, a Map becomes {}, and a circular reference throws.
This is the one place the difference is documented rather than inferred: IndexedDB Browser states plainly that sets, maps and typed arrays are not supported. IdxBeaver serializes through a versioned wire format specifically so these survive a round trip through an export and back. Kahuna documents the same care here, and goes further than we do on File and ImageData values.
IndexedDB is not the only storage you are debugging
A session bug is rarely confined to one storage API. The token is in a cookie, the feature flag is in LocalStorage, the cached response is in Cache Storage, and the user data is in IndexedDB. Four of these tools are IndexedDB-only; idb-crud adds LocalStorage and SessionStorage. IdxBeaver covers all five surfaces in one panel, so you are not switching tools mid-investigation.
When not to use IdxBeaver
A comparison page written by the author of one of the tools is worth exactly as much as its willingness to say this part.
- You want storage inspection without opening DevTools. IdxBeaver is a DevTools panel, so it only exists while DevTools is open. Kahuna’s overlay and idb-crud’s drawer both work outside DevTools, which is a genuinely nicer fit for a quick look at a flag. Kahuna additionally shows the database count for the current site on its toolbar icon, so you can tell a site uses IndexedDB without opening anything at all.
- You need to create a database, or add a store or index. IdxBeaver can delete a database or an object store but cannot create or alter one. Kahuna has a schema editor for exactly this.
- You want charts or store size breakdowns. IndexedDB Explorer advertises size calculations and chart generation. IdxBeaver does neither.
- You are on Firefox or Safari. IdxBeaver is Chromium 120 or newer today. Kahuna ships on both Chromium and Firefox, so it is the answer for a mixed-browser team.
- You need Service Worker debugging.That belongs to Chrome’s own Application panel and IdxBeaver does not try to replace it. The comparison against the built-in panel covers where each one wins.
- You want the most widely installed option. IdxBeaver is new. Several of these tools have been around considerably longer.
Which one should you pick
- Editing a handful of values. Any of them. IndexedDBEdit is the smallest and most focused.
- Browsing a small store in a nice table. IndexedDB Browser, which documents its scope honestly and does that job well.
- Quick look without DevTools. Kahuna or idb-crud.
- Editing a schema, or working on Firefox. Kahuna.
- Large stores, real queries, exports, unusual types, or more than one storage API. IdxBeaver.
