MILabs
Tools

Typesense search for business websites and documents

Typesense is an open-source search engine a business can run itself or use hosted, set up by MILabs for fast, typo-tolerant search.

Updated 2026-02-22

What Typesense is

Typesense search is an open-source engine that gives a business fast, forgiving search across its own records. It indexes records such as product listings, help articles, documents or contact details, and returns matching results quickly, even when a query contains a typo. Results can be filtered by fields such as category, price range or date. The engine is queried over HTTP, so a website or an application can send a search and display the results. Typesense can run on a server that a business controls, which keeps its index in the business's own environment, or it can be used as a hosted service run by the vendor. The project supports vector search, so records can be matched by meaning as well as by keyword. Official client libraries are available for common programming languages. A business does not need a search specialist to begin, though a large index benefits from careful schema design. Typesense is a search layer rather than a chat assistant, most useful where a business already holds information that people need to find quickly.

Typesense: official site

Typesense in plain language

A search index is a separate copy of information, arranged so that a query can be answered in a fraction of a second. A database stores records and answers precise questions about them. A search engine asks a different question: which records are most likely to be what this person meant? That difference explains why a shop that filters a product table by exact name often gets poor results from a shopper who types a word slightly wrong.

Typesense works by reading each record, breaking its text and fields into searchable terms, and storing those terms in a structure built for lookup. When a query arrives, the engine compares the words against that structure, allows for small spelling differences, and ranks the matches using rules the business can adjust. Fields that describe a record, such as a category or a date, can be marked so that they narrow results without affecting the ranking of the text itself.

The index is a derived copy. The original catalogue or document store remains the source of truth, so a change made there has to reach the index before it appears in search. Deciding how often that happens and what happens when it fails is part of the work that makes search reliable.

What businesses use it for

  • Helping shoppers locate products by name, size or colour on an online store.
  • Retrieving internal help articles, policies and procedures by keyword.
  • Locating past quotes, contracts or case files across a shared document folder.
  • Running a site search box that corrects typing mistakes.
  • Pairing enquiries with the right service page or team member.

Who it suits

  • Online shops with large catalogues that need quick, forgiving site search.
  • Professional firms holding many documents that staff consult every week.
  • Organisations that would rather keep their search index on their own infrastructure.

How MILabs sets it up

MILabs starts by talking with the owner or team lead about what the search should cover and who will use it. Together you choose whether the index runs on a server the business controls or on a hosted plan, and that choice is written down with its reasons. We then connect the systems the business already has, such as the product catalogue, the website, the document store or the help desk, so records reach the index as they change. Permissions are set so staff see only the records they are allowed to see, and the website's search key is limited to read access. Before handover, we agree with the owner who looks after the index, who adds new collections and who responds if results stop updating.

Builds it fits

Industries it fits

Typesense in practice

In a business that uses Typesense, most of the visible work sits on the website or in an internal tool, where a search box sends queries and shows results. Behind it, someone decides which collections exist, such as products, articles or files, and what each record looks like in the index. A product listing might carry a name, a category and a stock flag, while a document record might carry a title, a matter number and a date. These choices shape what people can filter by, so they are best made with the staff who know the records.

Once the schema is settled, updates become a steady rhythm. New items are added when created, changed items are re-sent when their details change, and removed items are taken out so customers never see a discontinued line. Anyone who maintains the index also logs changes to fields, since a renamed field can quietly break a filter the website depends on.

Typo tolerance forgives a shopper who misspells a word, but it cannot make a category spelled three different ways appear as one. Teams usually spend part of the early effort standardising category names, removing duplicate listings and filling gaps in descriptions, because that cleanup decides whether filters behave sensibly.

good questions, straight answers

Frequently asked questions

How does MILabs set up Typesense?
MILabs begins with the owner or team lead, agreeing which records the search should cover and whether the index runs on a server the business controls or on a hosted plan. We connect the existing catalogue, website or document store, set permissions so staff see only what they should, and agree in writing who maintains the index and who responds when results fall behind. Work outside Sydney is delivered remotely.
Where does Typesense store data?
Typesense keeps its search index on the server where it runs. A self-hosted index stays on a machine that the business or its IT provider controls, while a hosted index sits on infrastructure run by the vendor. The original records remain in the systems they came from, such as the catalogue or document folder, and Typesense holds a searchable copy. The owner confirms the choice during setup, before any records are sent.
When is Typesense the wrong choice?
Typesense is the wrong choice when a business needs a conversational chatbot, because it returns search results rather than written answers. It is also a weaker fit where the data is small enough that existing site search or a shared folder search already does the job, or where nobody can keep the index current as records change. In those cases, MILabs may suggest a different approach and records the reasons in writing.
Can my team use Typesense without training?
Yes, staff who only search use the front end of a search box and need no training. Anyone who adds collections, changes fields or reads the index needs a short walkthrough and a written note on how the schema is organised. MILabs provides that handover during setup and agrees with the owner who carries these tasks afterwards. Teams without technical staff usually hand those tasks to an existing IT provider.
Can MILabs build on top of Typesense?
Yes. MILabs can build a search layer on Typesense that connects to a website, an internal tool or an assistant. Typical work includes sending records to the index, defining fields for filtering, and wiring a search box or an assistant's retrieval step to the index. The owner agrees the scope in writing before work starts, and that document sets out what each build covers.
let's make something happen

Workflow automation: how we can help

Choose how you want to work on this with MILabs. Both options start with an enquiry.

Discovery is free, no obligation from the first chat.

Let us build it for you

We build it and look after it.

Build it yourself, we'll help

You build it with our direction, teaching and a weekly check-in.