June 22, 2026
AI pic, instant comment chaos
Manticore Search 27.1.5: Auth, sharding, conversational and faster vector search
Search tool gets passwords, smarter chat, and speed boosts as commenters roast the AI art
TLDR: Manticore Search added major upgrades like built-in security, chatbot-style search, and faster performance, making the software easier and safer to use. But in the comments, the biggest reaction wasn’t the tech—it was one reader instantly roasting the AI-themed artwork and refusing to read further.
Manticore Search just dropped a big update, and on paper it’s a serious glow-up: built-in logins and permissions, easier data splitting for bigger workloads, chatbot-style question answering, and faster handling of those modern “find similar stuff” searches. Translation for non-search-engine people: this tool wants to be more secure, more helpful, and much faster without forcing users to duct-tape extra services around it.
But the real fireworks? The community reaction was gloriously blunt. Instead of diving into the long list of upgrades, one commenter came in with the kind of drive-by review that can hijack an entire tech thread: “AI splash image; didn’t read”. That tiny grenade says a lot about the mood right now. Even when a product ships genuinely useful features, some readers are so exhausted by anything that looks like AI branding that the cover image alone becomes the story. It’s part protest, part meme, part internet survival instinct.
That creates the delicious drama here: Manticore is adding practical tools people actually ask for—security, speed, easier setup, and chat over your own documents—but the vibe check gets derailed by anti-AI aesthetics almost instantly. The joke is that the release is packed with real infrastructure upgrades, yet the loudest reaction is basically, “nice picture, I’m out.” In classic comment-section fashion, the software shipped features; the audience shipped a mood.
Key Points
- •Manticore Search 27.1.5 introduces built-in authentication and authorization, sharded tables, conversational search, faster HNSW builds, and faceting improvements.
- •The release notes cover changes shipped from versions 25.0.1 through 27.1.5 and include fixes across KNN, replication, and protocol compatibility.
- •Version 27.0.0 adds built-in auth/authz with staged upgrade requirements, matching auth data across daemons, and JOIN CLUSTER behavior that replaces local auth data.
- •Version 26.0.0 changes replication storage layout so replicated tables use the normal data_dir/<table> path, which may require migration or re-synchronization in some setups.
- •Vector-search improvements include KNN optimizations, local ONNX embeddings support, faster ONNX inference, and multithreaded HNSW builds that reduce build and rebuild time for large vector tables.