Scales efficient language processing with open-source accessibility.
357M visits/mo+1%#2 Coding & Development
Valori
Vector database with deterministic state and graph support for auditable AI systems
$0 /monthStarting price
About Valori
Valori provides a vector database designed for reproducible state in AI applications. It stores, indexes, and searches vector data through a REST API, ensuring that the same inputs produce identical outputs across different machines. The system uses fixed-point (Q16.16) computation instead of floating point to guarantee deterministic results. Every write operation commits to a BLAKE3-chained log and returns a receipt containing pre- and post-write state roots, enabling verifiable and replayable state recovery. The platform also supports graph relationships alongside vector similarity search, allowing users to query both data types within a single system. Valori is accessed primarily via a REST API with official Python and TypeScript SDKs, requiring no local infrastructure to begin use. It is positioned for applications where knowing the exact state behind a result is critical, such as retrieval-augmented generation, recommendation systems, and AI agent memory. The product emphasizes auditable, cryptographically verifiable infrastructure for AI workloads that demand predictability and compliance.
Key features
- Vector similarity search
- Graph relationship queries
- Cryptographically verifiable state
- BLAKE3-chained log for state tracking
- REST API with Python and TypeScript SDKs
- Fixed-point (Q16.16) computation
- State receipts for pre- and post-write verification
- Multi-node cluster replication (Pro/Enterprise)
Use cases
- Retrieval-augmented generation with verifiable context
- Recommendation systems using similarity-based ranking
- AI agent memory with persistent, queryable state
Pros
- Deterministic computation with fixed-point math for reproducible results
- Cryptographically verifiable state with BLAKE3 receipts for auditing
- Combines vector similarity search with graph relationship queries
- REST API with official Python and TypeScript SDKs
- Open source components and public documentation
Cons
- Public beta only with limited project and record capacity
- Pro and Enterprise tiers require private preview access
- No mention of multi-region deployment or global availability
- Fixed-point math may limit floating-point precision in some use cases
Frequently asked questions about Valori
What is Valori and what does it do?
Valori is a deterministic vector infrastructure with built-in graph support designed for AI applications that require reproducible and auditable state. It stores, indexes, and searches vector data through a REST API, ensuring the same inputs produce identical outputs across different machines using fixed-point computation.
Who should use Valori?
Valori is positioned for applications where knowing the exact state behind a result is critical, such as retrieval-augmented generation, recommendation systems, AI agent memory, and data analysis. It suits industries like ecommerce, legal, finance, healthcare, and high-frequency trading where verifiable vector search matters.
How does Valori ensure deterministic and verifiable state?
Valori uses fixed-point (Q16.16) computation instead of floating point to guarantee identical outputs on any machine. Every write operation commits to a BLAKE3-chained log and returns a receipt containing pre- and post-write state roots, enabling verifiable and replayable state recovery.
What integrations or SDKs does Valori support?
Valori provides a REST API and official Python and TypeScript SDKs for easy integration. It requires no local infrastructure to begin use, allowing users to authenticate and search vectors directly via HTTPS requests or SDKs.
What are the main limitations of Valori?
Valori is currently in public beta, with Pro and Enterprise tiers in private preview, which may limit immediate access to advanced features. The system is designed for deterministic and verifiable state, which may introduce performance trade-offs compared to traditional floating-point vector databases.
How do I get started with Valori?
To get started, create a project on Valori Cloud, obtain an API key, and begin storing vectors or running searches via the REST API or official SDKs. The platform offers documentation and an API reference to guide users through the process.
