AI Tools
Weaviate
Open-source vector database with keyword and vector retrieval.
What it is
Weaviate is an open-source vector database used to store and search data embeddings. It combines vector similarity search with keyword (BM25) search in a single hybrid query, which helps retrieval stay accurate when embeddings alone miss context. It offers modules for vectorizing data at import time, a GraphQL and REST API, multi-tenancy, and can be self-hosted or run on its cloud service. Teams use it for retrieval-augmented generation, semantic search and recommendation systems.
Weaviate pricing
Weaviate does not publish a plan ladder we could verify. Free tier available.
Prices read from the vendor's own pricing page on 2026-08-24. Vendors change plans without notice — check the source before you budget.
Key features
Hybrid search
Runs BM25 and vector search in parallel and combines their scores.
Fusion controls
Supports relative-score or ranked fusion plus an alpha weighting parameter.
Managed or self-hosted
Offers Weaviate Cloud and Docker or Kubernetes deployment paths.
Backup and restore
Supports cloud-object-storage backups, selected collections and incremental backups.
Strengths and trade-offs
What works well
- Hybrid retrieval: Combines BM25 keyword and vector results in one query.
- Deployment choice: Use Weaviate Cloud or run the same database yourself.
- Permissive license: The database repository uses the BSD 3-Clause license.
- Configurable fusion: Adjust keyword/vector weighting and result fusion behavior.
Where it falls short
- Self-hosting responsibility: Production security, scaling, storage, monitoring and backups require configuration.
- Cloud billing lifecycle: Paid clusters continue billing until they are deleted.
- Version boundaries: Client examples target current releases and preview features may change.
- Backup caveats: Restore compatibility and multi-tenant backup behavior depend on version and tenant state.
Who it is for
Teams that need combined keyword and semantic retrieval and want a choice between managed cloud and self-managed infrastructure.
Pick it or skip it
Pick Weaviate if
- Search systems that benefit from tuning keyword and vector relevance together
- Organizations that require an open-source self-hosting option
- Teams willing to operate Kubernetes or pay for a managed cloud cluster
Skip it if
- Your search needs are satisfied by a small embedded or direct in-process index
- You cannot own self-hosted production operations or accept cloud usage billing
- Your backup plan has not been tested against the exact Weaviate and tenant configuration
Our verdict
Choose Weaviate when hybrid ranking controls and deployment choice matter; validate relevance tuning and operational requirements on your own dataset.
Frequently asked questions
- Can Weaviate run locally for development?
- Yes. Besides Docker and Kubernetes deployments and the cloud service, Weaviate can run embedded inside a Python application.
- Does Weaviate support hybrid search?
- Yes. Vector similarity and BM25 keyword search combine in one hybrid query with configurable weighting.
- Does Weaviate have a free plan?
- Yes. Weaviate offers a free tier with limits; paid plans unlock higher usage or team features.
- Who is Weaviate best for?
- Teams that need combined keyword and semantic retrieval and want a choice between managed cloud and self-managed infrastructure.
- What are the main drawbacks of Weaviate?
- The trade-offs we noted: Self-hosting responsibility: Production security, scaling, storage, monitoring and backups require configuration.; Cloud billing lifecycle: Paid clusters continue billing until they are deleted.; Version boundaries: Client examples target current releases and preview features may change..
- When should you not use Weaviate?
- Skip Weaviate if: Your search needs are satisfied by a small embedded or direct in-process index; You cannot own self-hosted production operations or accept cloud usage billing; Your backup plan has not been tested against the exact Weaviate and tenant configuration.
What we checked
- Public APIOffers a documented API you can build against.Yes
- Mobile appHas a native app for iOS or Android, not just a mobile website.No
- Open source / self-hostableSource is open and the tool can be run on your own infrastructure.Yes
- SSO (SAML)Supports SAML single sign-on on at least one plan.Yes
“Not checked” means exactly that — we have not verified it, and we do not guess.
Evidence and freshness
Status: Official sources reviewedReviewed: 2026-08-24
- Weaviate Docs — Hybrid search ↗
Checked 2026-08-24 · Supports: BM25 and vector combination, parallel retrieval, fusion methods, alpha weighting, threshold limitations
- Weaviate Docs — Installation guides ↗
Checked 2026-08-24 · Supports: managed cloud, self-managed database, Docker and Kubernetes paths, Windows limitation, release-candidate warning
- Weaviate — Official repository ↗
Checked 2026-08-24 · Supports: open-source database, BSD 3-Clause license, objects and vectors, deployment options
- Weaviate Docs — Backups ↗
Checked 2026-08-24 · Supports: backup providers, selected collections, incremental backups, restore version caveat, multi-tenant limitations
- Weaviate — Pricing (automated check) ↗
Checked 2026-09-03 · Supports: An automated check confirmed this vendor pricing page was reachable on the date shown. No plan name, price or limit was extracted from it.
Limit: Vendor documentation was reviewed; ToolCompare did not independently benchmark retrieval quality, memory use, scaling behavior or cloud cost. The pricing source was added by an automated check: the page was fetched on this date and contained pricing information. The other sources on this record were read by an editor.
Tools that do the same job as Weaviate
Same job, different pricing. Verified starting price first.
More AI Tools tools: see the full AI Tools category