Features

Better results. Less waiting. Less work.

Everything you expect of a modern search engine, and the parts Indx does differently. Every search feature is in the free version.

  1. 01Better resultsPeople find what they meant, even when they typed it wrong.
  2. 02Less waitingResults on every keystroke, and changes searchable at once.
  3. 03Less workNo analyzers to tune, and an AI that tells you what to fix.

Better results

Finds what people mean

Results people see as relevant, not only ranked correctly. Up to three typos in a word, and pattern matching that keeps guessing past that. Every form of a word from one search. And the title that says it before the description that mentions it, in any language, with nothing to configure first.

  1. 1HonningmelonWhole word
  2. 2HoninmelonTypo
  3. 3Honning melonSplit
  4. 4Melon honningSplit, turned

And, when the query allows part of it

  1. 5HonningPartial
  2. 6MelonPartial
Honningmelon is Norwegian for honeydew melon. Every way it was written, in one list, from one search.
  • Typo tolerance

    Up to three typos in a word, counted by edit distance. Past that, pattern matching keeps guessing, so even a badly mangled word can still find its record.

  • Every form in one query

    Whole words, the start of a word, misspelled words, and words written together or apart, in the same query. No second search, and no separate autocomplete.

  • Accents, one way

    Search for jalapeno and you get jalapeno and jalapeño alike. Type the ñ on purpose, and the records spelled that way come first. On by default.

  • Search as you type

    A full ranked search on every keystroke, with the word being typed read as the start of a longer one. No separate suggest endpoint.

  • Built-in relevancy ranking

    Search for titanic, and Titanic comes before A Night to Remember, which only mentions it in its description. A word in a short field like a title counts for more than the same word in a long text, without weights or tuning.

  • Field weights

    Say that a title counts for more than a description, for every search or for one.

  • Hybrid search

    Bring your own embeddings, and blend vector and text scores in one ranked list.

  • Any language

    No tokenizer, stemmer or stop words to set up per language. Compound words are split as the index is built. Irregular forms, like mouse and mice, take a synonym.

  • Synonyms

    Teach it the words your visitors use for yours, one way or both. They apply on the next search, with nothing re-indexed.

  • Your JSON as it is

    Nested objects by dot notation, and arrays matched on any of their elements.

Better results

Narrows the list

Pick Science fiction and the years 1977 to 1983, and every count on the page follows. The filters, facets and sorting a catalogue needs, counted over the whole dataset.

Science fictionAdventure1977 to 1983

Genres

Counted over the whole result, so the numbers and the list agree. Illustration with film data.
  • Filters

    Values, ranges and true or false, combined with AND, OR and NOT, over the whole dataset and not just the page.

  • Facets

    A count for every value, and the lowest and highest of a number, that always agree with the count beside the list.

  • Sorting

    A field orders equally good results, so the best match stays first. With no search text, the field is the order.

Less waiting

Fast, and always current

Every keystroke gets a full ranked search, so the list keeps up with the person typing. Change a price, and the next search already has it.

  1. s9,412 results0.31 ms
  2. st4,207 results0.24 ms
  3. sta2,516 results0.22 ms
  4. star1,833 results0.27 ms
A ranked list on every keystroke, with no separate suggestions. Illustration with film data.
  • Under a millisecond

    Embedded, a query runs in your own process in under a millisecond, with no network round trip and no cold start.

  • Real-time updates

    A change is searchable at once: no queue, no rebuild. A whole dataset can be replaced with no downtime.

  • Webhooks

    Changes in your CMS arrive within seconds. Sanity and Contentful need no code.

  • Hibernation

    A dataset can give its memory back and keep its documents, and be woken again when it is needed.

Less work

You stay in control

See that people search for “sofa” and choose what you call a couch, and add the synonym. Lift the summer range from June to August. All in the browser, with no code.

Summer rangeActive
Lifts
season is summer
Strength
From
1 June
Until
31 August
It starts and stops on its own dates; rules stack.
  • Console

    Upload and configure datasets, preview results, manage teams and check the engine, with no code.

  • Search statistics

    Searches, clicks and conversions over 7, 30 or 90 days: the top queries, the misspelled ones that still found something, the languages. Counted on your server.

  • Boost rules

    Lift what you want seen first by your own fields, with a strength, and dates to start and stop.

  • Personal boosting

    Boost lists per visitor or group, such as what they have bought before.

Less work

AI on both sides

Someone types “a film to watch with kids tonight”: the results stay, and an answer from your catalogue is offered under the field. For you, an agent reads the statistics each morning and proposes what to change.

The assistant can answer this↵
The results stay; the question is offered under the field. How readily is up to you, from strict to proactive.
  • Search or ask

    Decision LLMs offer a question when a phrase reads like one, in modes from strict to proactive. The answer agent works from your index, with any LLM: fewer tokens, and better questions than single words.

    ExperimentalDocs
  • Observr

    Each morning it reads your statistics and proposes synonyms, field weights and settings. Nothing changes until you say yes.

    Pro, early accessDocs
  • MCP server

    Every server has one. Agents can search your datasets, and never change them.

  • A skill for coding agents

    Your coding agent knows the library and the server, and builds the integration with you.

Less work

Runs where you want it

Add a NuGet package and search in your own process, or run the server on your own machines or in your own Azure subscription. Either way, your data stays with you.

dotnet add package IndxSearchLib --prerelease
var engine = new SearchEngine();
engine.Init(stream);
var title = engine.GetField("title")!;
title.WordMatching = true;
title.PatternMatching = PatternMatchingSetup.Default;
stream.Position = 0;
engine.Load(stream);
engine.Index();

var result = engine.Search(new Query("matrix", 10));
Embedded: no server, no network hop. The same engine runs as the Indx server.
  • IndxSearchLib

    The engine as a NuGet package, in your own process. No server to run, for a .NET app that owns its data.

  • Indx server

    An HTTP API for any language, and for several applications sharing one index.

    BetaDocs
  • React components

    Search field, results, filters, sorting and Search or ask, ready for React.

  • Self-hosted, GDPR friendly

    The library sends nothing anywhere. The server sends us none of your data, unless you turn on Observr.

  • Azure

    Deploy into your own Azure subscription, in the region you choose.

  • API keys by team

    Search only, read only or full access, for one dataset or all, with a lifetime you set.

  • Terminal monitor

    Each dataset’s state, the CPU, the memory and the searches a minute, in your terminal.

Library or server

The same engine, two ways to run it

Embed it for a single .NET app that owns its data. Host it for any language, and for several applications sharing one index.

FeatureLibraryServer
Typo tolerance, ranking, filters, facets, sorting, hybrid search, synonyms
Yes
Yes
Runs in your own .NET process, under a millisecond
Yes
No
HTTP API for any language
No
Yes
React components
No
Yes
Console in the browser
No
Yes
Search statistics
No
Yes
Scheduled boost rules
No
Yes
Webhooks
No
Yes
API keys by team
No
Yes
MCP server and terminal monitor
No
Yes
Search or ask, and Observr
No
Yes

Try it on your own data