Weaviate Reviews & Product Details

I like being able to keep vectors and application data living together in one place. Generative Search is especially interesting to me when retrieval is embedded directly inside an AI workflow. Once the index and query patterns are tuned properly, performance stays predictable and consistent. The integrations also fit nicely around the AI application stack I’m working with. I appreciate that the vector index configuration gives me real engineering knobs to turn, rather than feeling like a black box. Multi-tenancy aligns well with the SaaS architecture I work with, and filtering combined with vector retrieval feels like a natural, practical pairing. Review collected by and hosted on G2.com.

Bad schema decisions tend to stick around. Debugging vectors is a different beast than debugging SQL, and it takes a different mindset and tooling. Pricing also needs to be grounded in workload math, not based on a generic ROI assumption. Review collected by and hosted on G2.com.

I really loved Weaviate's native hybrid search feature. It's perfect for handling complex queries in our insurance claim documents, where adjusters often need to use alphanumeric policy codes with natural language questions. This feature combines BM25 keyword scoring with HNSW vector similarity out of the box, which is fantastic. Another aspect I like is the recent upgrade to the Python client (V4 API), especially the type hinting, which integrates seamlessly with our Python codebase. This has made development much smoother and easier for us. I also appreciate Weaviate's ability to manage both data residency and security constraints effectively and the option to tune infrastructure to control storage costs. Review collected by and hosted on G2.com.

In our project, dialing in the hybrid search requires quite a bit of effort and manual tuning, particularly when adjusting the alpha parameter (keyword vs vector) and going through trial and error to get the fusion ranking correct. Additionally, while Weaviate Cloud offers ease of use, managing the open-source version locally via Docker presents noticeable operational complexity compared to a fully managed serverless database. Review collected by and hosted on G2.com.

I like that Weaviate makes it fairly straightforward to work with vector data and search through it. The documentation is useful when setting things up, and the dashboard makes it easy to keep track of the projects and collections. It also works well for experimenting with AI and search-related use cases. Review collected by and hosted on G2.com.

It can take some time to understand the different concepts if you're new to vector databases. Some of the configuration options can also feel a little overwhelming at first. Review collected by and hosted on G2.com.

What I like most about Weaviate is how straightforward it makes working with vector search and embeddings. The setup feels intuitive, and it’s easy to store, search, and retrieve relevant data for AI applications without adding unnecessary complexity to my workflow. Review collected by and hosted on G2.com.

What I dislike about Weaviate is that some of its advanced features and configuration options can feel a bit overwhelming at first. It can take a while to understand the different settings and to get everything configured exactly the way you want. Review collected by and hosted on G2.com.

Weaviate is an all-in-one platform for building vector search, RAG, and agent memory management for us. With it, we design, build, and ship the entire AI stack, from local development through to our AWS production environment. Above all, it’s open source, which helps eliminate vendor lock-in concerns for our organization and also provides the flexibility to customize, improving overall performance.

With its vector database, we store, index, and retrieve different types of media information for our products, which supports scaling AI agentic systems. It also includes an Explorer that runs semantic, keyword, and hybrid search with aggregation, without needing to write GraphQL that saves time for our engineers.

Weaviate embeddings also help deliver efficient, faster models like Snowflake designed for enterprise-level retrieval operations. Review collected by and hosted on G2.com.

Although it has decent features, the UI feels slightly outdated based on my experience. When it comes to integrations with third-party platforms outside of the machine learning ecosystem, it’s not quite there yet. That said, because of the open-source community on GitHub, I expect the number of integrations to grow over time for other tech stacks that we use daily.

It does offer a 7-day free trial, but after that, calculating the overall monthly cost is complicated. It charges different services at different rates, which makes it hard to understand the final price per month. We have to use a calculator to add up all the final costs. Review collected by and hosted on G2.com.

The biggest advantage of this Weaviate platform is there built in a I power search and its retrieval augmented generation applications without making the task much complicated right from the vector database infrastructure. They also manage the cloud service, which makes the deployment very quick while featuring the semantic search and hybrid search, along with the automatic vectorisation and integration with popular llm providing the development effort reduction significantly. Their documentation is also comprehensive, and their app is very well designed for both prototyping and production deployment. Review collected by and hosted on G2.com.

Their advanced concepts, such as schema design and sharding, along with cluster optimization required some learning before we could fully leverage on our platform. For very large data sets and cloud costs, this can increase the more built-in monitoring, visualisation, and cluster management capabilities directly within the cloud console. Also, their pricing uh structure should be made transparent and given a quick clarity right from the onboarding stage. Review collected by and hosted on G2.com.

It natively supports hybrid search, allowing developers to combine dense vector search (for semantic meaning) with traditional keyword search (BM25). This is widely considered the gold standard for retrieving highly accurate context in RAG applications. Review collected by and hosted on G2.com.

Instead of using standard SQL, Weaviate's primary query language is a custom implementation of GraphQL. While powerful for complex graph relationships, developers accustomed to traditional relational databases often report a steep learning curve. Review collected by and hosted on G2.com.

What stands out most about weaviate is its native hybrid search and multi-tenancy capabilities, which make combining BM25 keyword matching with dense vector search effortless. It allows us to deliver ultra-fast semantic video discovery and personalized viewer recommendations across huge OTT metadata catalogs. Review collected by and hosted on G2.com.

Setting up self hosted clusters and tuning HNSW index memory parmeters for large scale video catalogs requires significant infrastructure overhead. Additionally, breaking SDK changes between major version updates can require unexpected maintenance for our automated OTT metadata ingestion pipelines. Review collected by and hosted on G2.com.

I use Weaviate for efficiently storing vector data, which is crucial for embeddings and the retrieval phase of a RAG implementation. I appreciate its open-source nature and the efficient vector retrieval, which is necessary for ensuring the retrieval phase is accurate and fast. Additionally, AWS support is great, and the initial setup was straightforward. Review collected by and hosted on G2.com.

I think finetuning the self-hosting option is a bit complex. The RAM usage can be high depending on the dataset, so if self-hosting, several decisions regarding this have to be made. You have to scale horizontally in some cases if you want to keep it running smoothly. The initial setup was great, but finetuning is not as easy. Review collected by and hosted on G2.com.

As the Email Marketing & Design Coordinator at Creative Edge Design Studio, the best part about Weaviate for my daily workflow is its powerful natural language semantic search, which completely fixed my biggest workflow headache with our old Pinecone vector database. Before switching, I could only search by exact keywords, which meant I could never quickly find tailored email templates for niche client projects. Last quarter, I needed to make a warm, minimalist re-engagement newsletter for a local boutique skincare brand, and I knew we had similar past drafts, but I couldn’t recall the exact file names or keywords. I wasted almost two hours manually digging through cloud folders and old Pinecone search results with no luck. With Weaviate, I can type plain descriptive sentences about tone, design style and client industry, and it pulls perfectly matched archived newsletters and marketing snippets instantly. I also love that I can build separate asset collections for retail, hospitality and creative clients, which keeps my campaign research hyper-targeted. It pairs extremely well with LlamaIndex, letting me generate brand-consistent email drafts only from our studio’s approved content, no generic AI copy. Since we’re a 32-person small design team with no data engineers, I’m grateful I can tweak basic metadata schemas and organize new campaign assets on my own. This tool has cut my email pre-production research time in half, letting me focus more on custom email layout design for Figma and polishing Mailchimp campaign content instead of hunting for reference materials. Review collected by and hosted on G2.com.

My biggest frustrations with Weaviate all stem from it being built for technical engineers, not marketing/design staff like me, and it creates constant avoidable work delays. The lack of a visual GUI for bulk content uploads is my top issue. A month ago, I tried archiving 60+ finished client newsletter CSV files to update our asset library. All the natural line breaks and soft paragraph spacing in my email copy broke the scripted batch uploads repeatedly. With no drag-and-drop tool to auto-fix formatting errors, I spent an entire afternoon manually editing every single file just to finish the import. On top of that, its hybrid metadata filtering is super unreliable in real work scenarios. When I was building a seasonal holiday email campaign for a local café client recently, I filtered my search strictly for hospitality industry assets. Even with precise tagging, Weaviate kept pulling unrelated retail brand promotion emails into my results. I had to manually sift through dozens of mismatched drafts, killing time I reserved for email layout design. To make it worse, there’s no auto incremental indexing for my LlamaIndex workflow. After I uploaded new hospitality email templates, I forgot to manually reindex, and LlamaIndex generated outdated, mismatched email copy for my client campaign. There’s also no native link to Mailchimp, forcing me to manually copy all finalized content over. These small but frequent flaws make Weaviate’s great core functionality unnecessarily cumbersome for non-technical creative teams. Review collected by and hosted on G2.com.