Upstash | Vector logo

Upstash | Vector

Embeddings 4.1/10 Jun 17, 2026

Serverless vector database for embeddings, semantic search, metadata filtering, namespaces, and RAG retrieval.

Card required
No

Upstash | Vector scores 4.1/10 for free value in Embeddings. The main constraint: Free tier: 10K daily query/update operations, 1 GB max data/metadata size, 1,536 max dimensions, and 100 namespaces.

Note: no FPS in parallel report — used audit recommendation
Category Embeddings Kind Managed vector database Caveats More primitive than Upstash | Search: you manage document chunking, embedding strategy, and retrieval shape. Use Search when you want a higher-level document search product. Renewal monthly Card required No Eligibility anyone

Free value score

4.1 / 10
Show the math
Overall 4.1/10
value = 4.1/100
weighted_components_v1
Free value 2/10
value = 0.3 $/mo
fixed_log_v1
Search fit 5.2/10
value = 5.2/100
policy_score_v1
Access 7.6/10
value = 7.6/100
policy_score_v1

Key metrics

1 tokens/mo Monthly tokens policy_default: Conservative scoring fallback: no renewable monthly token allowance is published in this item.
1 tokens/min Tokens/min policy_default: Conservative scoring fallback: no comparable token throughput allowance is published in this item.
1 GB Storage

What you get

10K query/update operations per day free

Dense, sparse, and hybrid vector search options

Metadata filtering and namespaces

Good fit for custom RAG indexes

Gotchas

  • 10K daily operations

    Free tier allows 10,000 query and update operations per day, resetting each day.

  • 1 GB total data/metadata

    The free index is hard-capped at 1 GB of total vector data and metadata storage, with no overflow.

  • 1 free database index

    Only one Vector index is available on the free plan; the FAQ notes up to 10 indexes free but the plan table shows '1 Free DB'.

  • Max 1,536 dimensions

    Free tier caps vector dimensions at 1,536, while paid tiers support up to 3,072 (Pay-as-you-go/Fixed) or 5,000 (Pro).

  • 100 namespaces max

    Free tier is limited to 100 namespaces; paid tiers allow 10,000 or unlimited namespaces.

  • No uptime SLA

    The free plan has no uptime SLA; paid plans offer 99.9% or 99.99% SLA.

  • Community support only

    Free tier receives only community support; email and dedicated Slack/Discord support require paid plans.

  • Scale to zero / shared infra

    Free indexes scale to zero when idle, causing cold-start latency on first request after inactivity.

  • Limited regions

    Free plan is available in N. Virginia (AWS), Ireland (AWS), and Iowa (GCP) only; region choice may not match user's geography.

  • Max 200M vectors x dimensions

    The free plan supports up to 200 million vectors-times-dimensions capacity, versus 2 billion on paid plans.

Sources

vector-db embeddings semantic-search rag serverless
Open service Source ← All Embeddings

Alternatives

Same category · similar quota
Voyage AIVoyage AI · finite_credit · Hosted embeddings + reranker API (free tokens)
200M tokens

Voyage AI (now part of MongoDB) offers state-of-the-art retrieval embeddings and rerankers, with a large free token allowance included on every account.

Checked 28d ago
One-time
Qdrant Cloud Free TierQdrant Cloud Free Tier · free · Managed vector database
4.7

Managed vector search database with a free-forever single-node cluster for prototypes, including free cloud inference on select models.

Checked 28d ago
One-time
Weaviate CloudWeaviate Cloud · freemium · Managed Vector Database
5.1

Fully managed, AI-native vector database with a forever-free Sandbox cluster for building search and RAG apps.

Checked 27d ago
One-time