Manticore Search 27.1.5 Adds Authentication, Sharding, Conversational Search, and Faster Vector Search

Manticore Search version 27.1.5 has been released, significantly strengthening the search engine's security, scalability, and support for AI workflows.

Key updates:

• Built-in Auth/Authz: Supports users, passwords, Bearer tokens, and fine-grained permission control across MySQL, HTTP/HTTPS, distributed remote agents, and replication operations.
• Sharded tables: Handles data distribution and lifecycle management across shards directly within the engine to facilitate large-scale, write-intensive deployments.
• Conversational search: Lets you ask questions directly on top of vectorized tables via the CREATE CHAT MODEL and CALL CHAT commands. This makes it easy to implement document Q&A and support assistants without building a separate retrieval layer.
• Vector search performance improvements: HNSW build speed has been improved, along with fixes related to KNN (K-Nearest Neighbors).
• Other improvements: Includes improvements to faceting and aggregations, as well as enhanced replication stability.