Meilisearch

Meilisearch v1.56: filters up to 10x faster with 'sampledCount' sorted facets. Plus, when your embedder is slow or down, small updates still go through and big jobs stop early. github.com/meilisearch/...

RAG and MCP get mixed up a lot, but they do pretty different things. Say a customer’s order never showed up and they got charged twice. RAG is how your assistant finds the refund policy. MCP is how it checks the order and issues the refund. www.meilisearch.com/blog/rag-vs-...

Our September newsletter is out, starting with Quentin de Quelen's IBC2026 talk: search, discovery, and recommendations from one index, demoed live. Plus what we shipped this summer. www.meilisearch.com/blog/septemb...

"Something to keep coffee hot on my commute" doesn't share a word with "insulated travel mug." Semantic search looks for meaning, so it still finds it. Searching for a SKU? That's where full-text search shines. Full comparison: www.meilisearch.com/blog/full-te...

Most RAG pipelines chop documents into chunks and hope the right one comes back. RAPTOR builds a tree instead: summaries at the top, detail at the bottom, so a query can enter at the level that holds the answer. www.meilisearch.com/blog/raptor-...

Already running Meilisearch and search feels slow? New blog helps to diagnose why. Performance trace in Search preview breaks a query into stages, and each stage points at something specific – filters, searchable attributes, response size. Full post: www.meilisearch.com/blog/diagnos...

Until now, disk performance was something you inferred. Now, the updated infrastructure tab shows your reads, writes, and throughput against your current limit. When you need more, you upgrade it yourself. Standard costs nothing extra. Higher tiers are paid.

Semantic and hybrid search used to mean wiring up your own embedding setup first. The new AI Embedders page lets you connect OpenAI, Voyage AI, Cohere, Mistral, Gemini, Cloudflare, or Ollama from any index. Prototype in minutes, keep the same setup in production. Open any index to try it.

Moving off OpenSearch? We compared 10 alternatives on features, pricing, and real user reviews: Meilisearch, Algolia, Bloomreach, Vespa, Typesense and more. Why teams switch, and how to pick the right one: www.meilisearch.com/blog/opensea...

Search doesn't have to feel like infrastructure archaeology. With Lucene you build faceting, typo tolerance, and ranking manually. We compared 10 alternatives on features, pricing, and integrations. www.meilisearch.com/blog/lucene-...

Meilisearch v1.53 is here 📣 Foreign filters (experimental) now work on sharded setups – filter one index with data from another, across multiple machines. Each filter can now retrieve up to 1,000 documents, 10x the previous limit. github.com/meilisearch/...

RAG systems split documents into chunks – split them wrong, and even the best LLM struggles to retrieve the right content. Split them correctly, and accuracy and relevance improve. We break down 13 chunking strategies 👉 www.meilisearch.com/blog/rag-chu...

@LogRocket refreshed their 2021 Meilisearch getting-started guide for 202. Everything from your first index to hybrid search, built-in RAG via /chats, recommendations, and pinning, all through one API: blog.logrocket.com/getting-star...

Meilisearch v1.51 is out 🎉 Search up to 5.4x faster on field-heavy documents (153 ms → 28 ms in one production workload), Dynamic search rules that activate based on filters. Details:https://github.com/meilisearch/meilisearch/releases/tag/v1.51.0

Meilisearch or Manticore - which is the better fit? In our latest guide, you’ll learn how Meilisearch and Manticore differ in architecture and performance, which is better for startups, SaaS apps, and large datasets, & key factors to consider. Full blog here:

Keeping a search index fresh shouldn't mean babysitting a cron script. The certified Meilisearch plugin for @kestra.io allows you to index into Meilisearch as a workflow task – logged, retried, scheduled, observable. No glue code. www.meilisearch.com/docs/getting...

LLMs hallucinate. In healthcare, that's a risk to patients. RAG grounds AI answers in trusted medical sources instead of letting the model guess. Our new guide on building medical RAG systems 👇 www.meilisearch.com/blog/rag-for-medical-data

Synonyms in Meilisearch are now up to 13x faster (v1.49.0). They load only when a word actually matches one, so large synonym lists no longer slow search down. github.com/meilisearch/...

Validating a document template used to be tedious. Meilisearch v1.48.0 adds an experimental render route: render any template or fragment and see the result, before and after you configure an embedder. Turn on the renderRoute experimental feature to try it: cloud.meilisearch.com/login

Feeding an LLM your whole knowledge base can be slow and costly. Context distillation passes only the relevant bits – fewer tokens, lower cost, faster answers. How it works:

Search personalization now works with federated search 🎉 You can tailor results to each user, even across multiple indexes in a single search. This feature is an add-on; reach out to support@meilisearch.com and we'll get you set up!

VivaTech starts this week 🇫🇷 We'll be at our booth. Come grab a coffee with the team and talk search + AI, or book a time to chat. ☕ See you there!

Fast search, fast company. Meilisearch just landed at #7 on the @sifted.eu 100: France & Benelux. Proud to grow alongside the most ambitious startups in the region! Full ranking: sifted.eu/intelligence... #Sifted100 #SiftedLeaderboards

CarJager runs 1.3M+ car searches a month with no servers to manage. Moving from Algolia to Meilisearch: no sorting limits, a bill that makes sense, and a search that scales itself during peaks. Read how👇

VivaTech turns 10, and we'll be right in the middle of it. 17-20 June, Paris. Come see what fast, AI-powered search and retrieval feels like: faster discovery, higher conversions, happier customers, in milliseconds. Hall 7.1, Stand 1H42-003. We'll save you a demo. #VivaTech2026

Second time at VivaTech for the Meilisearch team 🇫🇷 June 17–20 in Paris. Come find the team, see what we're building, and talk search + AI. We're at Hall 7.1 · Stand 1H42-003. See you!

Is your search actually showing people what they're looking for? Our new guide on search relevance covers the metrics that matter, how to measure them, and the pitfalls to avoid. 👉️

RAG without guardrails is like launching a rocket with no navigation – exciting, but risky. Our new post covers who needs them, how Meilisearch helps, and why skipping them is a bad idea. 👉 www.meilisearch.com/blog/rag-gua...

RAG sounds magical – until you try to make it work with your actual data tables. What helps, where teams get stuck, and a few examples that clicked: meilisearch.com/blog/rag-for-structured-data

That's a wrap on Launch Week 🎉 Mon: Sharding & Replication Tue: SSO/SAML + MFA Wed: Chat UI Thu: Search Performance Inspector Fri: Document Joins + Dynamic Search Rules Catch up on everything here: [https://www.meilisearch.com/launch-week-q1-2026 ]

Closing out Launch Week with two betas: 🖇️ Document Joins: link indexes like SQL tables. Search results come back enriched with related data. 📌 Search Rules: pin documents to exact positions, define your own conditions you define. Try both now, free. Full release:

Chat shipped in Meilisearch last year. We've since put a lot of work into improving the setup experience, and it's out now. Gillian, who built and shipped this, put together a walkthrough showing off the new setup. Find it in the Launch Week page:

Day 2 of Meilisearch Launch Week 🚀 Enterprise SSO/SAML is here: SAML 2.0 + OIDC, works with Okta, Azure AD, Google Workspace. Also: MFA is now free for all Cloud accounts. Enable it today. More tomorrow 👀 www.meilisearch.com/launch-week-...

Meilisearch Launch Week is live 🔥 Day 1: Sharding & Replication - one of our biggest infrastructure releases ever. Scale beyond a single node. Instant failover. Global routing. And they're fully composable. Learn more: www.meilisearch.com/blog/shardin...

Pre-weekend reminder: Meilisearch Launch Week starts next Monday, April 13. Multiple features dropping in a row. At least two of them have been a long time coming. See you Monday 🙌

Three things dropping in Meilisearch Launch Week next week: – Horizontal scaling – Enterprise security (SSO + MFA) – AI chat, straight from your index ...and more. We kick off Monday 👀 www.meilisearch.com/launch-week-...

Meilisearch Launch Week starts next Monday April 13 Five features. Three themes: infrastructure scale, enterprise security, developer experience. First drop Monday - sign up to stay updated 👀 www.meilisearch.com/launch-week-...

RAG-as-a-Service keeps coming up in AI conversations – but what does it actually mean? We wrote a guide covering what it is, how it works, why companies are choosing it, and what to look for before picking a solution. 👉 www.meilisearch.com/blog/rag-as-...

Meilisearch Launch Week is coming next week: 📅 April 13th–18th. Register your interest to get early access and updates: 👉️ share-eu1.hsforms.com/1zljrJkBuTG2...

Automotive inventory search is uniquely hard, and buyers filter across a dozen attributes at once. Here's how Touchless solves it with Meilisearch: www.meilisearch.com/blog/touchle...

You know what's gotten better at checking its own work? 👀 AI. Self-RAG lets models evaluate their own outputs and decide when to go fetch more context. Our latest article breaks down how it works 👉 meilisearch.com/blog/self-rag

You've asked an AI something - and it confidently got it wrong. RAG fixes that. It pulls in real data before answering, so your AI stops guessing and starts being useful. Build better RAG pipelines with Meilisearch 👉 meilisearch.com/blog/rag-in-ai

Samtec inc runs 1.5M monthly searches across a catalog with trillions of part combinations. Engineers don't browse - they search for specifics, and slow or irrelevant results cost sales. Read the full story on our blog 👇 www.meilisearch.com/blog/samtec-... #Search #B2B #Ecommerce

Our CTO bsky.app/profile/kero... just dropped another one of his deep-dive posts. We patched LMDB – the database engine inside Meilisearch – to make our vector store 3x faster. 6 embeddings/sec → 20, on a real customer with 19M vectors. It's a good one: www.meilisearch.com/blog/3xfaste...

We're heading to KubeCon + CloudNativeCon in Amsterdam (March 23–26)! 🇳🇱 Building search into your cloud native stack? Reach out to set up the time, or come find Chiara from our team at the event. #KubeCon #CloudNativeCon

Custom search infra is costly, complex & slow to scale. More teams are switching to Search-as-a-Service — enterprise-grade search through simple APIs. ✅ Lower costs ✅ AI-powered relevance & personalization ✅ Faster UX Is it right for you? 👉 meilisearch.com/blog/search-as-a-service

Most search indexes end up bloated - you copy the same data everywhere just to get complete results. Our new 'foreignKeys' links indexes together. Store a reference, get back the full object at search time. Your indexes stay lean. Your results stay rich. Full release:

The March Meilisearch update is here! New ranking rules, replicated sharding in the engine, major embedding performance gains, and a bunch of Cloud UI improvements. Oh, and we officially joined the Rust Foundation 🦀 Full update: