AACSearch vs Algolia
How AACSearch compares to Algolia on retrieval, merchandising, personalization, pricing model and self-hosting.
Algolia popularized instant, typo-tolerant search-as-you-type as a hosted API, and it remains a strong, mature product with a large ecosystem. AACSearch takes the same "hosted search API" shape but bundles a broader platform — merchandising, personalization, recommendations, experiments and a quality lab — as first-party features rather than a growing list of separate add-ons, and prices AI features (the assistant, voice and image search) as metered credits rather than a fixed platform tier.
Switching engines is the main thing holding teams back, so AACSearch ships a working migration tool, not just a promise: paste your Algolia config export into the console's migration wizard and it maps your searchable/facet attributes, synonyms and rules into a new collection, then bulk-imports your records — sign up free and it's waiting for you at Dashboard → Migrate from Algolia.
Retrieval
Both platforms support typo-tolerant full-text search, faceting, and fast response times on indexed data. AACSearch additionally ships vector and hybrid search, natural-language query understanding, and a zero-hit rescue ladder (layout correction, transliteration, query broadening) built in on every collection — useful if your traffic includes mixed-script queries (Cyrillic/Latin, for example).
Merchandising and rules
Algolia's Rules engine and AACSearch's search rules cover similar ground — pins, boosts, redirects, filters — with a difference in how campaigns and scheduling are modeled: AACSearch's campaigns support recurrence (a rule that's only active on weekends, or during business hours) and versioning with one-click rollback out of the box.
Personalization and recommendations
Algolia's Personalization and Recommend are separate products with their own pricing. AACSearch bundles behavioural personalization and 16 recommendation model types (similar items, bought-together, session-based, sequential, graph-based, and more) into the core platform.
Pricing model
Algolia prices primarily on search operations and records. AACSearch's request-based plans (Free, Starter, Pro) work similarly, but AI-heavy operations — the assistant, embeddings, voice and image search — are billed separately from a credit wallet, so a catalogue that does a lot of plain search but little AI doesn't subsidize one that does the opposite.
| Feature | AACSearch | Algolia |
|---|---|---|
| Typo-tolerant full-text search | Yes | Yes |
| Vector & hybrid search | Yes | Add-on (NeuralSearch) |
| Merchandising rules | Yes | Yes (Rules) |
| Personalization | Included | Separate product |
| Recommendations | 16 model types included | Separate product (Recommend) |
| A/B experiments | Included, with confidence intervals | Add-on (A/B Testing) |
| Offline relevance evaluation (nDCG/MRR/MAP) | Included (Quality Lab) | Not built in |
| AI assistant (RAG with citations) | Included | Not built in |
| 1C-Bitrix module | Yes | No |
| Self-hosted option | No — hosted platform | No |
Frequently Asked Questions
Is AACSearch a drop-in replacement for Algolia?
Not literally — the request/response shape and query syntax differ. Migrating means re-pointing your search calls at AACSearch's API and re-indexing your catalogue, which is typically a day or two of work for a mid-sized catalogue. The console's migration wizard handles the schema, synonym and rule mapping from your Algolia config export automatically; API keys/ACLs and stopword lists are a short manual step afterward, and rules with more than one condition or Algolia-specific features (optionalFilters, replica indices, Query Suggestions) need a quick manual check — the wizard's review step lists exactly what it skipped before you apply anything.
Does AACSearch support the same faceting and filtering Algolia does?
Yes — facet counts, hierarchical facets, numeric and date ranges, and combined filter expressions are all supported.
Which one is cheaper?
It depends on your traffic mix. A catalogue with heavy AI usage (the assistant, voice, image search) will likely find AACSearch's metered-credit model more predictable; a catalogue that's pure text search with no AI features may find either platform's base pricing comparable — compare the pricing page against your expected request volume directly.