Turbopuffer alternatives
When your vector bill scares you and much of the corpus is cold, Turbopuffer is the interesting answer.
Overall 8.21 (curated_estimate, high) · score asOf 2026-09-14 · last reviewed 2026-09-14. This is not an Elo and not a live price board. Provider · JSON
Jobs: rag-memory, cheap-vector-scale, billion-scale-vector, embedded-vector, vector-object-hybrid-memory
| Provider | Why listed | Overall | Δ vs subject | asOf |
|---|---|---|---|---|
| Supabase (pgvector) Depot default for most agent apps under ~25–50M vectors. Don't add Pinecone until Postgres actually hurts. | same_jtbd_pick, same_jtbd_runner · rag-memory, cheap-vector-scale | 8.64high | +0.43 | 2026-09-14 |
| Zilliz (Milvus) When you're truly at massive scale, Milvus/Zilliz earns its complexity. Don't start here for a chatbot MVP. | same_jtbd_runner, same_jtbd_pick · cheap-vector-scale, billion-scale-vector | 8.16high | -0.05 | 2026-09-14 |
| LanceDB Great embedded/object-store vector layer. Pair with Mem0/Zep only if you need higher-level memory APIs. | same_jtbd_pick · embedded-vector | 7.78medium | -0.43 | 2026-09-14 |
| Qdrant Depot pick for serious vector workloads with filters. Self-host or Cloud — either way you can leave. | same_jtbd_runner · rag-memory, cheap-vector-scale, billion-scale-vector | 8.59high | +0.38 | 2026-10-19 |
| Pinecone Still the zero-ops vector default. If you already run Postgres, try pgvector before paying Pinecone rents. | same_jtbd_runner · rag-memory, billion-scale-vector, vector-object-hybrid-memory | 7.95high | -0.26 | 2026-09-14 |
| Upstash Vector Prototype/edge RAG peer when you already on Upstash. Prefer Qdrant/Pinecone for serious filtered production… | same_jtbd_runner · rag-memory, cheap-vector-scale | 8.15high | -0.06 | 2026-09-14 |
| Cloudflare R2 Depot pick when agent artifacts get downloaded a lot. S3-compatible without the egress grief. | same_jtbd_runner · vector-object-hybrid-memory | 8.75high | +0.54 | 2026-10-19 |
| Amazon S3 Default object store gravity well. Consider R2/Cloudflare or DO Spaces when egress dominates. | same_jtbd_runner · vector-object-hybrid-memory | 8.6high | +0.39 | 2026-09-14 |
| Upstash Redis Default serverless-edge-redis. Use classic Redis Cloud when you need always-on fat memory / Redis modules… | same_jtbd_runner · rag-memory | 8.49high | +0.28 | 2026-09-14 |
| Tigris R2 remains Depot object-storage / cheap-egress default (CDN+Workers gravity). Tigris when you want global… | same_jtbd_runner · vector-object-hybrid-memory | 8.15high | -0.06 | 2026-09-14 |
| Redis Use as agent hot memory/cache, not as your only 100M-vector archive. | same_jtbd_runner · rag-memory | 8.07high | -0.14 | 2026-09-14 |
| DuckDB (vector/extensions) Use for local analytics + light vector experiments. Production agent memory → pgvector/Qdrant/Mem0. | same_jtbd_runner · embedded-vector | 8high | -0.21 | 2026-09-14 |
Prices are not ranked on this page. For a job, use the sourced quote (separate input and output rows, Depot Pick unchanged): rag-memory, cheap-vector-scale, billion-scale-vector, embedded-vector.
Alternatives are other catalog providers that share a JTBD as pick or runner-up. If fewer than three exist, same-category active providers fill the list and are labeled same_category. Overall deltas are catalog scores. kind=curated_estimate is not a lab measurement. Pricing rows are sourced or explicitly null. Do not parse pricingNotes. A single normalized USD figure is set only for a provider with exactly one token row.