turbopuffer

used in

stacks

Serverless vector + full-text search with object-storage backend.

commercial usage https://turbopuffer.com ↗

Pricing scales with usage, not provisioned capacity. Strong fit for agent retrieval where read patterns are bursty and per-namespace isolation matters.

Strengths

Object-storage-backed; cheap at rest
Hybrid vector + BM25 search
Per-namespace isolation

Tradeoffs

Newer player; smaller ecosystem than Pinecone

Used in 13 stacks

Reference: Deep Browser Research Agent reference

research-deep

Reference: Team-Scale Coding Assistant reference

coding-assistant

Reference: Contract Extraction Pipeline reference

document-processing

Reference: Contract Redline Assistant reference

legal-review

Reference: Brand-Voice Content Generation reference

marketing-content

Reference: Meeting Assistant reference

meeting-assistant

Reference: Incident Response Agent reference

ops-incident-response

Reference: Mid-Scale RAG Over Internal Docs reference

rag-over-docs

Reference: Candidate Sourcing Agent reference

recruiting-sourcing

Reference: Financial Deep Research Agent reference

research-deep

Reference: Outbound SDR Research + Drafting Agent reference

sales-sdr-outbound

Reference: Mid-Volume Support Triage Agent reference

customer-support-triage

Reference: Inbound Voice Support Agent reference

voice-support

Get the weekly digest

New stacks, new tools, what's moving. No spam.