AACSearch vs Typesense Cloud
Typesense Cloud hosts a search engine; AACSearch is a search platform — tenant isolation, merchandising, personalization, quality lab and billing behind one OpenAPI contract. Here's how they compare.
Typesense Cloud gives you a hosted retrieval engine and stops there: schemas, typo tolerance, faceting, filtering and vector search, with everything above that layer left to your team. AACSearch is the layer above — a control plane with its own tenant API that adds isolation, merchandising, personalization, quality measurement, an AI assistant and billing on top of typo-tolerant, faceted, vector-capable retrieval. So the honest framing isn't "which engine is faster", it's "how much of a search product do you want to build yourself".
What's comparable
Typo-tolerant full-text search, faceting, filtering, sorting and vector/hybrid search are table stakes on both sides. If you know Typesense Cloud's query vocabulary — query_by, filter_by, sort_by, num_typos — AACSearch's search parameters will look familiar, and migration tooling imports existing collections and dumps.
What AACSearch adds
Tenant isolation. Typesense Cloud gives you a cluster; multiple applications sharing one cluster manage their own namespacing. AACSearch namespaces every resource per organization automatically, at the API layer, so a search from one tenant can never see another tenant's documents even when they share infrastructure.
Merchandising and rules. Pins, boosts, exclusions, redirects, banners and scheduled campaigns are not part of Typesense Cloud's API — AACSearch computes them on every search.
Personalization and recommendations. Behavioural profiles and 16 recommendation model types, built from your own event stream, with no separate service to run.
Quality lab and experiments. Golden query sets, offline relevance scoring, and A/B experiments with statistical confidence — none of which Typesense Cloud provides.
AI assistant, voice and image search. A retrieval-augmented assistant with citations, speech-to-query, and image-to-product search, metered from a credit wallet.
Billing and multi-tenancy. If you're building a product on top of search (a SaaS platform, a marketplace, an agency serving many client storefronts), AACSearch's per-organization API keys, quotas and billing remove the need to build that layer yourself.
| Feature | AACSearch | Typesense Cloud |
|---|---|---|
| Typo-tolerant full-text & vector search | Yes | Yes |
| Multi-tenant isolation | Built in | You build it |
| Merchandising rules | Built in | Not part of the service |
| Personalization & recommendations | Built in | Not part of the service |
| Quality lab (nDCG/MRR/MAP) | Built in | Not part of the service |
| A/B experiments | Built in | Not part of the service |
| AI assistant with citations | Built in | Not part of the service |
| Billing / API keys per customer | Built in | You build it |
| Collection schema & typo-tolerance control | Full, via the tenant API | Full |
Frequently Asked Questions
Should I use AACSearch or Typesense Cloud?
If you need only raw retrieval and are comfortable building tenant isolation, merchandising, and billing yourself, a hosted engine is a reasonable choice. If you want those layers built and maintained, AACSearch is the platform with them already in place.
Can I bring Typesense-style search parameters to AACSearch?
Yes — collection schemas, query_by, filter_by, sort_by, num_typos and vector search parameters follow the same shape in AACSearch's tenant API, so existing queries and integrations carry over with minimal changes.
Does AACSearch lock me into one engine forever?
The tenant API is AACSearch's own OpenAPI contract, with SDKs generated from it for TypeScript, PHP, Python and Go — not a pass-through of any engine's raw API. That is what makes the merchandising, personalization and multi-tenancy layers possible, and what keeps your integration stable regardless of what runs underneath.