Customers
Explore how ambitious teams are building with turbopuffer, and what it feels like to puff with the best.
Notion searches every page
pwtb ‘26core
SaaS
- $ saved annually
- millions
- documents
- 10B+
- write peaks
- 1GB/s
- namespaces
- 10M+
Legora searches every case
pwtb ‘26core
BYOC
- documents
- 2B+
- p90 latency
- 100ms
- recall@10
- 98%
Linear searches every issue
- cost reduction
- 70%
- documents
- 2B+
- p50 latency
- 13ms
- namespaces
- 4M+
Atlassian searches every wiki page
- search quality increase
- 19%
- p90 latency
- 60ms
- of documents
- billions
- recall@10
- 96%
Superhuman searches every email
- years of emails
- 1 → 5
- p90 latency
- 60ms
- cost reduction
- 20%+
- documents
- 10B+
fal searches every media asset
- media embeddings
- 300M+
- namespaces
- 225K+
- p50 latency
- 17ms
Mem0 searches every agent memory
- agent memories
- 400M+
- lower latency
- 70x
- avg recall@10
- 97%
Kepler searches every financial document
- documents
- 100M+
- p90 latency
- 85ms
- lower cost
- 10x
Pylon searches every support ticket
- to production
- 3 days
- namespaces
- 50k+
- documents
- 600M+
- p90 latency
- 24ms
Readwise searches every highlight
- vector and full-text search data
- 8TB+
- namespaces
- 200k+
- documents
- 1B+
- faster embeddings
- 8x
PlayerZero searches every line of code
- faster retrieval and reasoning
- 18x
- documents
- 2B+
- namespaces
- 45k+
TELUS searches every knowledge base
- namespaces
- 300K+
- documents
- 200M+
- active users
- 57k+
- p99 latency
- <100ms
