Indx Core v5 is launched, and Indx v2 is in beta
A faster search engine, now released, and a redesigned cloud platform in beta. See what is new and how to try them.
We have been rebuilding two of the biggest pieces of Indx, and they are now ready to use. Indx Core v5 is the next generation of the engine class library, and Indx v2 is a complete redesign of the hosted platform.
Both are available today. Indx Core v5 is launched as a stable release, version 5.0.0, and Indx v2 is in beta.
The same pattern matching you know, now faster, leaner, and easier to integrate.
What is new in Indx Core v5
The v5 engine keeps the pattern matching that makes Indx resilient to messy input, and adds a set of long-requested capabilities.
- Dynamic inserts. Add documents to a live index without a rebuild.
- Hybrid search. Combine pattern matching with LLM embeddings for semantic relevance.
- Refreshed 16-bit scoring model. Finer-grained relevance scores across the board.
- Query-time field boosting. Weight fields per query instead of baking it into the index.
- Better agentic search. Improved support for AI agents querying Indx directly.
- Faster and leaner. Performance improvements with reduced memory consumption.
Indx v2
Indx v2 is a complete redesign of the hosted platform that takes you from raw API calls to a full control plane. Everything you need to run search in production now lives in one place.
- Visual index management. Configure indexing, field roles, and boost rules from the dashboard, with no redeploys.
- Team owned datasets. Invite your team and assign roles so the right people get the right access.
- Real time notifications. Stay informed as long running and dynamic operations complete.
- Native MCP support. Connect Indx to your AI agents and tools over the Model Context Protocol.
- Live search preview. Tune relevance and preview results against your real data, right in the dashboard.
- Hardened fault handling. A more resilient engine that recovers gracefully and keeps search online.
- Resource manager with hibernation. Configurable hibernation idles datasets when they go quiet and wakes them on demand to reclaim resources.
The auth model is simpler, the endpoints are more consistent, and the new team concept keeps your datasets organized as you scale. If you build against the hosted API, v2 is the surface to grow on.
Try it out
Indx v2 is in beta and open source on GitHub. The server, API, and tooling are there for you to inspect, fork, and self-host on Azure or any provider that runs .NET.
Prefer to embed the engine directly? Indx Core v5 is on NuGet as a stable release. Add the package and you are searching in minutes.
Indx v2 is still in beta, so your feedback shapes it. Tell us what works and what does not.