Menu
Menu
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
Keyword search
Self hosting
Managed hosting
Cohere models
PaLM models
OpenAI models
HuggingFace models
AI / Semantic /Vector search
Image search
Multilingual search
Question answering
Hybrid search
Recommendations
User events
Personalized search
This type of search uses a vector database and works by matching meanings and contexts instead of literal words.
For example, “something to sleep on” would return beds even if the results don’t contain the word “sleep”.
The advantages of this search is that it needs less fine-tuning than keyword search and it is multilingual (example : the spanish user typing in his own language would get matching english results) but is more ressource intensive so more costly.
The compatible search engines that use this type of search are Elasticsearch, Opensearch, Algolia and Weaviate.
This is the most common type of search available. It is the lightest and least expensive form of search but could need fine-tuning to deliver a flawless experience to users.
If you are interested in this type of search, you could use Apache Solr, Elasticsearch, Opensearch and Algolia.
All of these have hosting solutions that provide a working environment so that you can start faster.
Combine multiple search algorithms to improve the relevancy of search results.
Using WPSolr and Weaviate, you can utilize the full power of both keyword search and semantic search simulteanously by combining them.
You can learn more about it from the official Weaviate documentation.
This type of search delivers a personalized experience to the users by sending user events to the search engine.
It is continuously learning to ensure that all the users receive the most relevant results based on their past actions on the website.
Some of the providers of personalized search are Recombee, Algolia and Google retail.
This is not too dissimilar from personalized search but instead of suggesting content when users do a search, they receive relevant content passively whenever they are using the website.
They are present in virtually all professional e-commerce websites: if you can spot a section where there is written “Products you might like” there is a very high likelyhood that a recommendation engine is in use.
When recommendations are active, whenever a user does anything (click on a product, add to cart, etc…) an event is sent to the engine.
The engine will then learn about all the users’s habits using the great amount of stored events and recommend the most relevant products or articles.
The compatible recommendation engines are Recombee, Algolia and Google Retail.