Releases Index
turbopuffer

turbopuffer

$npx @buildinternet/releases get turbopuffer
Recently Shipped3 releases · updated Jun 16, 2026

An i8 vector type landed in June, cutting storage and query cost by 75% compared to f32 for quantization-aware models.

Smaller vectors, lower cost — The new i8 type targets models already trained for quantized inference, making high-throughput retrieval meaningfully cheaper without a quality trade-off.

Namespace branching and fuzzy search — May brought instant copy-on-write namespace cloning via namespace branching, alongside a Fuzzy filter for typo-tolerant string matching. A C# API client shipped alongside an updated word_v4 tokenizer that runs roughly 3× faster than its predecessor, and API keys can now be scoped to list-namespaces only.

Sparse vectors, cross-cloud copies, and faster writes — April added sparse vector search support and cross-provider namespace copies between GCP and AWS. AWS write latency dropped 2.5× through a faster commit cadence. Full-text query length limits rose to 8,192 characters, attributes per namespace expanded to 1,024, and three new regions opened in São Paulo, South Carolina, and Belgium. Namespace pinning arrived as a way to hold frequently queried data in cache at high QPS without overpaying for idle time.

CJK full-text search gets dedicated tokenizers

This release2 featuresNew capabilitiesAI-tallied from the release notes

Full-text search now supports Chinese, Japanese, and Korean with the smartcn, kuromoji, and nori tokenizers, plus a new whitespace tokenizer.

Read more ↗

🏗️ Faster storage engine coming soon: follow the turbopuffer v3 devlog for updates 🪓 Sharding is generally available: scale a single namespace up to 256TB 🥇 Native reranking with models like Jev and Voyage Rerank 3 [private beta] ⏭️ Paginate results with offset 👀 Monitor a…

Read more ↗

🫛 Embed without an API call: native embedding is generally available 🧬 Added support for 4 new embedding models 🗺️ Browse namespace documents and run queries in the dashboard 🔟 Store arbitrary binary data in turbopuffer with the bytes data type ☁️ First public Azure region,…

Read more ↗

🫛 Embed without an external API call using native embedding (private beta)

Read more ↗

🧮 i8 vector type for quantization-aware models (75% reduced storage and query cost compared to f32)

Read more ↗

May 2026

Rollup

🌿 Namespace branching: instant copy-on-write namespace cloning ✏️ Typo-tolerant string matching with the Fuzzy filter 🗺️ View namespace metadata in the dashboard ️⃣ C# API client 🔤 word_v4 tokenizer, 3x faster than word_v3 ⏱️ Make asynchronous requests to copy_from_namespace…

Read more ↗

📌 Pin a namespace to cache for lower cost at high QPS 🕸️ Support for sparse vector search ↔️ GCP <-> AWS namespace copies with copy_from_namespace 🔎 Search the turbopuffer docs (type cmd+K!) 🏃‍♂️ Faster commit cadence on AWS for 2.5x lower write latency 💗 Increased…

Read more ↗

🗂️ Multiple vectors per document now available for everyone 🔐 Audit logs with SIEM integration [opt-in, beta] ⚡️ Up to 30% faster AND queries 📄 Copy docs as markdown with token counts 3️⃣ Control term frequency's influence on BM25 scores with k3

Read more ↗

💵 Query pricing reduced by up to 94% 🔤 Regex index for much faster Regex, Glob, and IGlob filters 🍵 Up to 20% faster filtered FTS queries 💎 Use attribute values to influence full-text search ranking 📐 Boost result recency with distance ranking in full-text search 🗂️ Store…

Read more ↗

🔡 FTS v2: up to 20x faster full-text search, now live for everyone ☯️ Up to 26% faster FTS queries on high-frequency terms with dynamic bit set encoding 🔌 turbopuffer MCP Server [beta] 🏷️ Match documents on any token with ContainsAnyToken 🔐 Permissions guide for…

Read more ↗

🧱 Redesigned inverted index structure for faster full-text search queries 📤 New object storage-native indexing queue for up to 10x faster queue time 🔦 kNN exact search for 100% recall on filtered vector search queries 🪣 Return a max number of search results per attribute…

Read more ↗

🏎️ FTS v2: up to 20x faster full-text search [opt-in, beta] 🔑 copy_from_namespace can now encrypt with a different key into the destination ✈️ Cross-region, cross-org copy_from_namespace for testing, backups, branching ⬆️ Max limit.total raised from 1,200 to 10,000 ➕ Sum…

Read more ↗

⏫ Rank by filter for full-text search 🧩 patch_by_filter 🔘 []bool support 🏎️ Improved performance for order-by queries 👁️ View indexing state in metadata API 📚 Read replicas for scalable read throughput (opt-in) 🔐 Cross-region…

Read more ↗

🧮 ANN v3: query 100B+ vectors with p99 of 200ms [opt-in, beta] 🚀 5x object storage throughput for faster cold queries and indexing 🔍 Prefix queries for full-text search 💧 Disable backpressure for large scale ingestions 🔐 Org-level option to enforce private networking 💎…

Read more ↗

🟰 Eq operator for array attributes 🗂️ Grouped aggregates (facets) 🇰🇷 South Korea region 🇮🇳 India region 🔀 Any filter operators for array attributes (e.g. AnyLt, AnyLte, AnyGt, AnyGte)

Read more ↗

July 2025

Rollup

🇸🇬 Singapore region 🇨🇦 Canada region 🕵️‍♀️ Private Service Connect + PrivateLink support 🎈 float type 🕳️ exclude_attributes query parameter 🪢 Regex filter operator 📋 Listing namespaces is now consistent 💎 Ruby API client GA release

Read more ↗

June 2025

Rollup

👩🏽‍⚖️ Conditional writes 🔣 Multi-query API 📝 Contains and ContainsAny filter operators 🐍 Python async API client ☕ Java API client GA release (with improved ergonomics) 🦫 Go API client GA release 💸 Discount queries on large namespaces (80% discount after 32GB)

Read more ↗

May 2025

Rollup

🐡 turbopuffer is generally available 🎊 v2 query API (unifies vector and full-text ranking) ✌️ Count aggregate function 🦫 Go API client beta release ⏩ Up to 4x faster filtering and full-text search ranking

Read more ↗

🥳 v2 write API (includes patch support) 💾 Up to 33% reduction in p90 query latency by using Direct I/O for local SSD cache 🔼 Max operator for full-text search 🙅 Not filtering parameter ☀️ Warm cache endpoint ☁️ AWS us-east-2 region

Read more ↗