Fast, encrypted, and honest about it.
Most database benchmarks quietly run warm, with caches on and encryption off. We don't. This is one pure database, benchmarked with every cache disabled and AES-256 encryption on, at 10 million rows — plus an honest, capability-by-capability look at how it compares.
Cold = the first run from cold disk with all caches disabled; warm = a repeat of the same query. No document cache, no aggregate cache, no shared-scan cache — and encryption at rest stays on the whole time.
10M head to head
10 million rows, head to head.
Warm query times at 10 million rows — 2.5M each across four types — InventDB vs PostgreSQL 18. InventDB runs cache-free and with AES-256 encryption on; PostgreSQL runs unencrypted. Same categories, same data volume, same machine.
InventDB figures re-verified on current main, July 2026 — single-thread 25.8 s warm / 82.7 s cold, multi-thread 44.2 s warm — unchanged from the run below.
| Category | InventDB | PostgreSQL | InventDB advantage |
|---|---|---|---|
| Basic SELECT | 53 | 1,702 | 32.1× |
| Exact match | 87 | 7 | PG faster |
| Range | 76 | 7 | PG faster |
| LIKE | 484 | 54 | PG faster |
| Compound WHERE | 702 | 14 | PG faster |
| ORDER BY | 123 | 8,324 | 67.7× |
| GROUP BY | 634 | 7,334 | 11.6× |
| GROUP BY + WHERE | 16 | 1,902 | 118.9× |
| JOIN 2-table | 4,141 | 16,354 | 3.9× |
| JOIN 3-table | 13,035 | 63,948 | 4.9× |
| JOIN 4-table | 2,003 | 38,981 | 19.5× |
| Secondary table | 83 | 1,399 | 16.9× |
| Pagination | 110 | 422 | 3.8× |
| NULL handling | 72 | 250 | 3.5× |
| HAVING | 237 | 3,742 | 15.8× |
| Edge cases | 5 | 725 | 145.0× |
| Complex | 3,598 | 13,238 | 3.7× |
| DISTINCT | 359 | 5,885 | 16.4× |
| Total | 25,818 | 164,288 | 6.4× faster |
Being honest: PostgreSQL wins the tiny warm exact-match and range lookups — those hit a hot index in single-digit milliseconds. InventDB wins 14 of 18 categories and every heavy one: ORDER BY, GROUP BY, all three JOIN depths, HAVING, DISTINCT and the complex multi-stage queries — while carrying encryption PostgreSQL isn't.
| Category | InventDB | PostgreSQL | InventDB advantage |
|---|---|---|---|
| Basic SELECT | 39 | 5,430 | 139.2× |
| Exact match | 240 | 6 | PG faster |
| Range | 200 | 6 | PG faster |
| LIKE | 220 | 89 | PG faster |
| Compound WHERE | 946 | 11 | PG faster |
| ORDER BY | 262 | 23,339 | 89.1× |
| GROUP BY | 1,239 | 26,108 | 21.1× |
| GROUP BY + WHERE | 18 | 5,533 | 307.4× |
| JOIN 2-table | 6,921 | 48,805 | 7.1× |
| JOIN 3-table | 19,070 | 220,444 | 11.6× |
| JOIN 4-table | 5,341 | 158,815 | 29.7× |
| Secondary table | 82 | 3,161 | 38.5× |
| Pagination | 178 | 1,194 | 6.7× |
| NULL handling | 19 | 740 | 38.9× |
| HAVING | 427 | 10,093 | 23.6× |
| Edge cases | 176 | 2,401 | 13.6× |
| Complex | 7,610 | 30,463 | 4.0× |
| DISTINCT | 1,114 | 13,995 | 12.6× |
| Total | 44,102 | 550,633 | 12.5× faster |
Under concurrency the gap widens to 12.5×. InventDB stays near-flat while PostgreSQL's join and aggregation categories degrade sharply — the 3-table JOIN reaches 220 s and the 4-table JOIN 159 s per thread. PostgreSQL figures are drawn from the InventDB 10M benchmark harness (2026), run on the same dataset and machine.
How it compares
Every capability, side by side.
A complete built-in-capability comparison — not a performance shoot-out and not a pricing table. The question isn't "who's fastest," it's "what ships in the box, without bolting on extra services." Verified July 2026 against official docs and release notes.
| Capability | InventDB | PostgreSQL 18 | MongoDB 8 (Community) | SQLite |
|---|---|---|---|---|
| Data model | ||||
| Schema-free JSON documents | Built in | JSONB | Yes | JSON1 |
| SQL query language | Built in | Yes | — (MQL) | Yes |
| Nested objects & arrays | Built in | JSONB | Yes | JSON1 |
| Reliability | ||||
| Full ACID transactions | Built in | Yes | Yes | Yes |
| Write-ahead log (WAL) | Built in | Yes | Yes (journal) | Yes |
| Self-healing corruption recovery | Built in | — (manual) | — (manual) | — (manual) |
| Security & encryption | ||||
| Encryption at rest | Built in | — (disk/TDE) | Enterprise | Add-on (SEE) |
| Row-level value encryption | Built in | Add-on (pgcrypto) | Yes (client keys) | Add-on (SEE) |
| Encrypted, queryable indexes | Built in | — | Yes (Queryable Enc.) | — |
| Database-managed rotatable keys | Built in | — (external KMS) | — (customer-managed) | — |
| Role-based access control | Built in | Yes | Yes | — |
| Indexing & performance | ||||
| Automatic indexing | Built in | — (manual) | — (manual) | — (manual) |
| Built-in key-value cache | Built in | — | — | — |
| Search | ||||
| Exact-match search | Built in | Yes | Yes | Yes |
| Full-text (keyword) search | Built in | Yes | Yes | Yes (FTS5) |
| Fuzzy / typo-tolerant search | Built in | Add-on (pg_trgm) | Add-on (Atlas Search) | — |
| Semantic / vector search | Built in | Extension (pgvector) | Add-on (mongot) | Extension (sqlite-vec) |
| Files & documents | ||||
| Encrypted file storage | Built in | — | — (GridFS unencrypted) | — |
| Text extraction, 25+ formats | Built in | — | — | — |
| Built-in OCR (image/PDF → text) | Built in | — | — | — |
| File version history | Built in | — | — | — |
| Built-in AI agent | ||||
| Natural-language Q&A | Built in | Add-on (SQL-gen only) | — | — |
| Multi-step analysis & reports | Built in | — | — | — |
| Natural-language data edits | Built in | — | — | — |
| Scheduled reports | Built in | — | — | — |
| Operations & deployment | ||||
| Single binary, zero dependencies | Built in | — (server) | — (server + mongot) | Yes |
| Driverless HTTP / REST API | Built in | Add-on (PostgREST) | Add-on (Data API) | — |
| Cross-platform (Win/Linux/macOS) | Yes | Yes | Yes | Yes |
Verified July 2026 against official docs and release notes; capabilities move fast — confirm with each vendor.
Methodology
How we ran it.
The setup, in full
Windows Server 2025, encryption on, all caches disabled. We ran the suite both single-thread and with 4 threads (one per core). Cold is the first execution from cold disk; warm is a repeat of the same query. The dataset is 10M records; the suite is 143 queries across 18 categories. Third-party capability data is drawn from public sources, July 2026 — we don't run competitor benchmarks here, only compare documented, built-in capabilities.
Run your own numbers.
Spin up your own encrypted database, load your data, and benchmark it however you like. Same SQL, same API, from a few thousand rows to a quarter-billion.