Skip to content

v2.0 beta 1

Choose a tag to compare

View all tags
@a-nt a-nt released this 08 Oct 16:57
· 398 commits to main since this release
v2.0-beta
5f4e9d1

Indx 2.0 beta 1

Self-hosted search for your JSON data, with a console for the people who run it.

Indx 2.0 is a new server around a new engine. Upload a JSON array, choose which fields are searchable, and you have typo-tolerant search over HTTP, a dashboard your whole team can use, and an MCP endpoint your AI agents can query. It replaces IndxCloudApi v1, and the repository is now called Indx.

indx-beta-cut.mp4

  Change things while they run

Replace every document in a dataset, or change how its fields are searched, and the new index is built beside the live one. Search keeps answering the whole time, and the switch happens when the new index is ready. Nobody has to plan a maintenance window.

  See what people meant

Indx matches on patterns in what people type, so it finds the marble lamp from "mrbl". The statistics catch it: which misspelled searches still found something, and what people chose. Alongside that you get the month at a glance, top queries, top documents and what people filter by. Statistics are experimental in this beta, and the numbers may still change.

  Work on it together

Datasets belong to teams, and everyone gets a role: Admin, Editor or Viewer. Switch team or dataset from the breadcrumb in one click.

Also new

  Native MCP
Point Claude, Cursor or any MCP client at your server with an API key, and your agents can search your data. 		  Hybrid search
Pattern matching and embeddings in one search. 		  Dynamic data
Insert, update and delete documents without a rebuild, by key or by filter.
  Live search preview
Tune relevance against your real data, with facets and filters, right in the console. 		  Boost rules and synonyms
Scheduled boosts for campaigns, and synonyms per dataset (experimental). 		  Hibernation
Quiet datasets sleep and wake on demand, so memory goes where the traffic is.
  Scoped API keys
Search, Read or Full keys, bound to a team and its datasets. 		  Faster and leaner
A new engine underneath, IndxSearchLib 5, with lower memory use. 		  Open source
MIT licensed. Run it anywhere .NET 10 runs.

Quick start

git clone https://github.com/indxSearch/Indx
cd Indx
dotnet run

Open https://localhost:5001. The first visit walks you through a short setup: create the admin account, name your team, and pick a few instance settings. It takes under a minute, and nothing needs configuring beforehand.

Requires the .NET 10 SDK.

Upgrading from v1

2.0 is not a drop-in upgrade: routes are now team-scoped, errors have a standard format, and the binary is renamed. The Indx Search skill ships a migration guide, and a coding agent using it can do most of the work. The full release notes list every change.

Companion projects

Licence

The server is MIT licensed. The bundled search engine is free up to 100,000 documents, and a free licence from the Indx License Portal removes the limit. The server can fetch it for you.

This is a beta

This is the first of a series of betas on the way to 2.0, each a numbered release here. Tell us what works and what does not: open an issue, and your feedback shapes the release.

Assets 2
Loading