01
Choose a self-contained file
Use a backup exported by the source application. A live database may keep recent changes in a separate WAL file, which this tool does not merge. The file information shows its size before you run further queries.
Database & SQL tools
SQLite viewerInspect a database without changing it
Inspect a database without changing it
SQLite viewer
Database and SQL
No database loaded.
Structure and indexes
Choose an object
Indexes
Query result
No database loaded.
Database files and SQL stay in this browser. Nothing is uploaded or saved as history.
Read-only · 50 MiB file · 20,000 SQL characters · 10,000 rows · 256 columns · 2,000,000 result units · 5 s execution / 20 s engine load.
100 rows per page; cells preview 200 characters. Exports include all loaded values. JSON: columns + rows, large integers as strings, BLOB as hex objects. CSV: formula protection, no lossless type round trip.
Open a database or try the example.
Use this SQLite viewer when an app backup arrives as a database file and you need to inspect its records before writing a script. Open SQLite files directly, then select a table or query a view.
A SQLite database viewer reveals structure as well as values. This SQLite browser keeps the query beside the result, so you can check which filter produced the rows you are about to export.
Why use this SQLite viewer?
Read the schema before choosing a query
The SQLite viewer lists tables and views, with the selected object’s CREATE SQL and its indexes. You can check a column name or unique index before writing a join; the original file is never rewritten.
Export the loaded result, across every page
A page contains at most 100 rows, but exports include every loaded row. A visible truncation message distinguishes a 10,000-row sample from a complete answer; changing SQL disables old exports.
Start with a consistent snapshot
02
Select a table or write a filter
Selecting a table runs a quoted SELECT. Editing SQL waits for Run query or Ctrl/Cmd+Enter. Use WHERE to narrow records and ORDER BY for a defined order; result pages do not sort the whole database.
03
Check the result boundary
Read the loaded-row count and truncation notice before you export SQLite to CSV. Add a stronger filter if the result was capped. JSON preserves column order and duplicate column names through separate columns and rows arrays.
Schema, queries and inspectable values
Real SQLite execution
The bundled sql.js WebAssembly engine evaluates a single SELECT, including supported joins, aggregates and SELECT-based CTEs. Work runs in a terminable Worker; a short but expensive query can be cancelled without blocking the page.
Tables, views and index definitions
The object selector includes tables and views. The structure panel shows the selected CREATE statement and its indexes, including automatic indexes without their own CREATE SQL. Schema information is read from the opened file.
Exports with explicit value handling
JSON retains NULL, represents integers outside the safe JavaScript range as decimal strings, and marks binary fields as objects with type and hexadecimal data. CSV quotes every field and prefixes formula-like text with an apostrophe for spreadsheet use.
What a read-only session can and cannot do
- Single-file, single-statement scope
- Only SELECT or a WITH query ending in SELECT is accepted. Writes, user PRAGMA, EXPLAIN, transactions, bound parameters, multiple statements and attached databases are rejected. Reserved write words used as identifiers must be quoted.
- Fixed resource budgets
- A file may be at most 50 MiB. SQL is limited to 20,000 characters; results to 10,000 rows, 256 columns and 2,000,000 serialized UTF-16 units. The row cap is marked as truncated; wider or larger results fail. SQLite heap allocation is capped at 128 MiB.
- Time limits and fresh results
- Engine loading has a 20-second budget and each database operation has five seconds. Cancel or timeout terminates the Worker. Editing SQL immediately makes an earlier result stale; run again before copying or downloading.
- Snapshots, not recovery or decryption
- SQLCipher, encrypted files, separate WAL/journal recovery and damaged-database repair are outside scope. No server connections, remote URLs, persistent query history or database editing are offered.
- Results are not database backups
- CSV and JSON contain the loaded query result, not all tables or their constraints. Cells preview only 200 characters; exports contain full accepted values. CSV loses NULL/type distinctions and changes formula-like text; use JSON when those distinctions matter. Text with invalid UTF-8 is rejected; query hex(column) to inspect its bytes.
SQLite viewer questions
What file types can I open?
Choose a standard SQLite 3 database, commonly named .db, .sqlite or .sqlite3. The header determines whether the file is accepted, not its extension. A SQL text dump, ZIP archive or encrypted database is not a directly readable file.
Can I edit the database?
No. Queries run against an in-memory copy with SQLite query-only mode enabled, and the interface rejects write statements. There is no save-database command. Use a database editor when you need to change records or schema.
What is the maximum file size?
The file limit is 50 MiB, equal to 52,428,800 bytes. Available browser memory can still cause a smaller file to fail. Large joins or large individual values can exceed query limits even when the input file is small.
Can I run any SQL command?
No. Use one SELECT or a SELECT-based CTE. SQL syntax is interpreted by the bundled SQLite version, and tables requiring unavailable extensions may not work. Error details remain visible below the panel.
Can I export tables to Excel or CSV?
You can download CSV for the loaded result and open it in a spreadsheet. This is not an XLSX export. Values beginning like spreadsheet formulas receive an apostrophe, and an explicit notice identifies any row truncation.
Found a query or file problem?
Inspect another database
Use the SQLite viewer to check the structure, refine a query and take only the result you have reviewed.