turbopuffer
@turbopuffer.bsky.social
search every byte 🔍 {vector, full-text} search engine built from first principles on object storage. 10x cheaper, scales to 100B. powers Notion, Cursor, Linear
a database is nothing without its guarantees
as part of our recent patch_by_filter launch, we clarified the ACID properties users can expect from turbopuffer
read more here: turbopuffer.com/docs/guaran...
as part of our recent patch_by_filter launch, we clarified the ACID properties users can expect from turbopuffer
read more here: turbopuffer.com/docs/guaran...
Guarantees
serverless vector and full-text search built from first principles on object storage: fast, 10x cheaper, and extremely scalable
turbopuffer.com
October 28, 2025 at 3:00 PM
a database is nothing without its guarantees
as part of our recent patch_by_filter launch, we clarified the ACID properties users can expect from turbopuffer
read more here: turbopuffer.com/docs/guaran...
as part of our recent patch_by_filter launch, we clarified the ACID properties users can expect from turbopuffer
read more here: turbopuffer.com/docs/guaran...
new: `patch_by_filter` updates all documents that match the provided filter, making bulk updates a breeze!
October 27, 2025 at 3:00 PM
new: `patch_by_filter` updates all documents that match the provided filter, making bulk updates a breeze!
new: rank by filter! boost scores when docs match a condition (e.g. scale > 1pib). plugs straight into rank_by, and works alongside full-text search
October 24, 2025 at 3:00 PM
new: rank by filter! boost scores when docs match a condition (e.g. scale > 1pib). plugs straight into rank_by, and works alongside full-text search
October 15, 2025 at 4:28 PM
tpuf ANN v3 can search 100 billion vectors with a p99 of 200ms
simplicity scales
[v3 in beta, unfiltered search, 1024D, k=10, 92% recall]
simplicity scales
[v3 in beta, unfiltered search, 1024D, k=10, 92% recall]
October 14, 2025 at 6:59 PM
tpuf ANN v3 can search 100 billion vectors with a p99 of 200ms
simplicity scales
[v3 in beta, unfiltered search, 1024D, k=10, 92% recall]
simplicity scales
[v3 in beta, unfiltered search, 1024D, k=10, 92% recall]
it was driving us mad that a 96 core box could only drive 13 Gbps from S3
turns out eBPF probes from our network monitoring imposed more overhead than anyone expected
5-10x improvement in network throughput → faster puffin'
turns out eBPF probes from our network monitoring imposed more overhead than anyone expected
5-10x improvement in network throughput → faster puffin'
October 13, 2025 at 3:00 PM
it was driving us mad that a 96 core box could only drive 13 Gbps from S3
turns out eBPF probes from our network monitoring imposed more overhead than anyone expected
5-10x improvement in network throughput → faster puffin'
turns out eBPF probes from our network monitoring imposed more overhead than anyone expected
5-10x improvement in network throughput → faster puffin'
September 16, 2025 at 11:07 PM
"Having our data in turbopuffer makes us feel ready for whatever comes next in AI-enabled software. It gives us the confidence to adapt quickly."
— Tristan Homsi, co-founder of Readwise
turbopuffer.com/customers/r...
— Tristan Homsi, co-founder of Readwise
turbopuffer.com/customers/r...
Readwise → turbopuffer, the use-case that inspired turbopuffer's creation
serverless vector and full-text search built from first principles on object storage: fast, 10x cheaper, and extremely scalable
turbopuffer.com
September 3, 2025 at 4:29 PM
"Having our data in turbopuffer makes us feel ready for whatever comes next in AI-enabled software. It gives us the confidence to adapt quickly."
— Tristan Homsi, co-founder of Readwise
turbopuffer.com/customers/r...
— Tristan Homsi, co-founder of Readwise
turbopuffer.com/customers/r...
new: native facet support with group_by
August 26, 2025 at 6:46 PM
new: native facet support with group_by
August 7, 2025 at 2:23 PM
every few months we re-run the landing page benchmarks against prod
vector search p90 down ~50%
vector search p90 down ~50%
August 6, 2025 at 3:00 PM
every few months we re-run the landing page benchmarks against prod
vector search p90 down ~50%
vector search p90 down ~50%
ripgrep on object storage?
August 5, 2025 at 8:37 PM
ripgrep on object storage?
our ruby client just dropped 💎
July 23, 2025 at 2:59 PM
our ruby client just dropped 💎
by popular demand, now introducing one-dimensional vectors 🎈
July 22, 2025 at 2:57 PM
by popular demand, now introducing one-dimensional vectors 🎈
"turbopuffer makes it too easy to build state-of-the-art AI apps" – Notion
turbopuffer.com/customers/n...
turbopuffer.com/customers/n...
Notion
serverless vector and full-text search built from first principles on object storage: fast, 10x cheaper, and extremely scalable
turbopuffer.com
July 17, 2025 at 3:16 PM
"turbopuffer makes it too easy to build state-of-the-art AI apps" – Notion
turbopuffer.com/customers/n...
turbopuffer.com/customers/n...
July 11, 2025 at 3:17 PM
July 9, 2025 at 3:16 PM
we reduced query pricing for large namespaces by up to 80%
puff more, pay less
puff more, pay less
July 7, 2025 at 8:24 PM
we reduced query pricing for large namespaces by up to 80%
puff more, pay less
puff more, pay less
S3’s conditional writes were the key ingredient that made our architecture possible
now it's our time to pay it forward
now it's our time to pay it forward
July 3, 2025 at 5:15 PM
S3’s conditional writes were the key ingredient that made our architecture possible
now it's our time to pay it forward
now it's our time to pay it forward
"Their responsiveness and shipping velocity make us feel like we are their only customer." – Tom Moor, Head of Engineering at Linear
turbopuffer.com/customers/l...
turbopuffer.com/customers/l...
June 26, 2025 at 3:20 PM
"Their responsiveness and shipping velocity make us feel like we are their only customer." – Tom Moor, Head of Engineering at Linear
turbopuffer.com/customers/l...
turbopuffer.com/customers/l...
June 6, 2025 at 5:53 PM
May 14, 2025 at 2:22 PM
May 12, 2025 at 4:00 AM
May 7, 2025 at 10:18 PM
modern NVMe drives are blazing fast, but software needs to be architected from the ground up to take full advantage of them
we recently rolled out a new I/O scheduler that uses Direct I/O to maximize performance of our SSD caches, avoiding page cache cpu and locking overhead
we recently rolled out a new I/O scheduler that uses Direct I/O to maximize performance of our SSD caches, avoiding page cache cpu and locking overhead
May 6, 2025 at 9:53 PM
modern NVMe drives are blazing fast, but software needs to be architected from the ground up to take full advantage of them
we recently rolled out a new I/O scheduler that uses Direct I/O to maximize performance of our SSD caches, avoiding page cache cpu and locking overhead
we recently rolled out a new I/O scheduler that uses Direct I/O to maximize performance of our SSD caches, avoiding page cache cpu and locking overhead