The use case: how Mem0 deploys turbopuffer
Mem0 runs its managed agent memory platform on turbopuffer, giving each customer a dedicated namespace for hybrid vector and full-text search with attribute filtering. Embeddings were migrated off pgvector in Postgres, which struggled with latency at scale.
More reported results
Figures published in turbopuffer’s case study. They are not independently measured by usedby.
97% average vector recall@10
70ms p90 hybrid retrieval latency on turbopuffer
150k+ customers, each with an isolated search index
Scaled from 100M → 400M+ memories without latency increase or recall drop
In the customer’s words
The biggest benefit of turbopuffer is the peace of mind that we won't fall over as we scale.Other companies using turbopuffer
Customers across industries, with public evidence.
See all turbopuffer customers
Go straight to the source
Based on turbopuffer’s published case study. We check the public evidence; the results are reported by the vendor.
turbopuffer.com- Last checked
- Nov 12, 2026
Open source capture What the story claims, and what we checked
What we compared with the page.
- The figure: 70xChecked
Printed word for word on the page, near the name of Mem0.
- The passage quoted aboveChecked
Copied word for word from the page, near the name of Mem0.
- Mem0 uses turbopufferChecked
Confirmed line.
- The result itselfNot checked
We quote it; we did not measure it.
Hybrid retrieval latency has consistently remained below 100ms — with tail latencies rarely exceeding 1s — even as Mem0 has scaled to 3TB+ of embeddings on turbopuffer.