Navigation
Sitemap
Every page published on the Vela site.
Product
- Home The managed semantic search engine: index your embeddings and query them by similarity in milliseconds.
- Features The engine itself: vector storage, HNSW index construction, query planning and metadata filtering.
- Pricing Plans, limits and usage-based pricing, from the free Discovery tier to billion-vector workloads.
- Integrations SDKs, frameworks and data sources that connect to the engine.
- Regions Where your vectors live: regions, data residency and latency.
- Compare How the engine compares with Pinecone, Weaviate, Qdrant and pgvector on the criteria that matter.
Developers
Engineering log
- All articles The full engineering-log index
- Why keyword search hits its limits What BM25 does well, what it cannot do, and why vector similarity search complements the lexical index rather than replacing it.
- Choosing your embedding dimension Memory cost per vector, how dimension affects recall, quantization and truncation: how to decide before indexing a large corpus.
- Reducing the p99 latency of a vector index Why the tail degrades before the median: efSearch, metadata filtering, reindexing, and how memory pressure shows up in high percentiles.
Company
- Security Encryption, isolation, compliance and how your vectors are protected.
- Customers How teams run semantic search in production.
- About The team, the thesis and how the engine came to be.
- Contact Talk to the team about an architecture, a quote or an incident.
- Legal notice Publisher, host, intellectual property, personal data and cookies.
- Sitemap This page
- Release 2.8.0 Filtering moves inside the index
- Release 2.7.0 Live reindexing
- Release 2.6.2 Tail latency under memory pressure