Home / Directory / Data & labeling / Vector / retrieval infra / turbopuffer
turbopuffer
Vector and full-text search database built on object storage, used by Notion, Linear, Atlassian, and Superhuman to search billions of documents.
AI and SaaS products that need vector and keyword search over very large, multi-tenant data where storage cost matters more than every query being hot.
Small apps with a few million vectors that already fit in Postgres, or workloads where every rarely used namespace must answer in milliseconds.
Verdict
turbopuffer is a search database company founded in 2023 by Simon Hørup Eskildsen, who spent almost a decade on Shopify's infrastructure team. Companies such as Notion, Linear, Atlassian, Superhuman, and Readwise use it for vector, full-text, and hybrid search, with data stored in object storage like S3 and cached on memory and SSD when queried. It matters because vector search grew expensive as AI products indexed every document, and this design lets large multi-tenant products search billions of records at lower cost.
Score Breakdown
How turbopuffer scores in the categories that matter to its buyers.
Pricing
turbopuffer bills usage against a monthly minimum. Every plan includes all database features.
| Plan | Price | What stands out |
|---|---|---|
| Launch | $16 / month minimum | SOC 2 report, GDPR-ready DPA, community Slack and email support |
| Scale | $256 / month minimum | Adds HIPAA-ready BAA, SSO, audit logs, private Slack channel |
| Enterprise | $4,096 / month minimum, 35% usage premium | Single-tenancy, BYOC, private networking, 99.95% uptime SLA, 24/7 support |
The Field at a Glance
How turbopuffer compares with other Vector / retrieval infra vendors we reviewed, by Overall Score and relative typical engagement cost.
turbopuffer scores 7.9 in Vector / retrieval infra, under Pinecone (8.3) and Zilliz (8.0), and ahead of Weaviate (7.6). Relative cost lands lower for this subcategory.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Search across billions of documents | Strong | Notion reports more than 10 billion documents. |
| Many small tenants | Strong | Linear runs more than 4 million namespaces. |
| Hybrid vector and keyword search | Strong | Vector, BM25 full-text, and hybrid queries are built in. |
| Bring your own cloud | Strong | Enterprise supports BYOC, as Legora and Atlassian use. |
| Cold, rarely used data | Mixed | Cold namespaces answer far slower than warm ones. |
| Tiny prototypes | Mixed | Works, but Postgres may be enough at small scale. |
Who it’s for
Good fit
- SaaS products adding AI search per customer
- Agent memory stores
- Code and document search at large scale
Poor fit
- Hobby apps with little data
- Workloads that cannot tolerate cold reads
- Teams that want a managed graph database
Review Excerpts
Below are excerpts from public reviews. Our team scoured public reviews, forums, and chat rooms to get a balanced view of customers' experience with this company. Paid reviews and pay-for-play sites such as Clutch were excluded.
“It's been fun to work with @turbopuffer on [the new search]. Admittedly keyword based search never really worked great for issues, and the results from embedding+FTS are now actually useful.”
“turbopuffer's economics have changed the way we think about building products that connect data to users and LLMs.”
On turbopuffer's own published benchmark of 10 million 1024-dimension vectors, a warm namespace answers in 14 ms at p50 while a cold namespace takes 874 ms, so rarely queried data is much slower on first hit.
Methodology
This page is an independent evaluation of turbopuffer for buyers comparing options in vector / retrieval infra. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. turbopuffer did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria sit under it in two groups.
- Buyer outcomes (for vector / retrieval infra)
- Cost at large scale
- Warm query latency
- Multi-tenant namespaces
- Cold query latency
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites such as Clutch are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the fit pattern described above.
Scope
turbopuffer is graded here as vector / retrieval infra. Criteria scores can move as more review volume and product checks are added.