Pricing
VecLabs pricing is built on two transparent, pay-as-you-go components: Shadow Drive storage and Solana transaction fees. There are no seats, no tiers, no monthly minimums.
How pricing works
VecLabs does not charge a platform fee during alpha. You pay only the underlying infrastructure costs directly:
| Component | Cost | Paid to |
|---|
| Vector storage | ~$0.05/GB/year | Shadow Drive |
| Merkle root post | $0.00025/write | Solana |
| Queries | Free | - |
| VecLabs platform fee | $0 (alpha) | - |
Storage cost calculator
Each vector is a fixed-size array of 32-bit floats (4 bytes per dimension) plus metadata overhead.
Cost per 1M vectors:
| Dimensions | Vector data | With metadata (~200B) | Monthly cost |
|---|
| 384 | ~1.46 GB | ~1.65 GB | ~$0.007 |
| 768 | ~2.93 GB | ~3.12 GB | ~$0.013 |
| 1536 | ~5.86 GB | ~6.05 GB | ~$0.025 |
| 3072 | ~11.7 GB | ~11.9 GB | ~$0.050 |
For 100K vectors at 1536 dimensions: ~$0.0025/month
Solana transaction costs
Every .upsert() call posts one Merkle root to Solana - one transaction regardless of how many vectors are in the batch. Batching is free.
| Write frequency | Monthly Solana cost |
|---|
| 100 writes/day | ~$0.75 |
| 1,000 writes/day | ~$7.50 |
| 10,000 writes/day | ~$75 |
| 100,000 writes/day | ~$750 |
To minimize Solana costs: batch your upserts. One .upsert() call with 1,000 vectors costs the same as one .upsert() call with 1 vector - $0.00025.
Comparison to alternatives
At 1M vectors, 1536 dimensions, 10M queries/month:
| VecLabs | Pinecone (s1) | Qdrant Cloud | Weaviate Cloud |
|---|
| Storage | ~$0.25 | $70 | $25+ | $25+ |
| Queries | Free | Included | Included | Included |
| Total/month | ~$0.25 | $70 | $25+ | $25+ |
| Data ownership | ✅ Your wallet | ❌ Their servers | ❌ Their servers | ❌ Their servers |
| Audit trail | ✅ On-chain | ❌ None | ❌ None | ❌ None |
| Encryption | ✅ Client-side | ❌ Server-side | ❌ Server-side | ❌ Server-side |
Pinecone and Qdrant pricing above reflects their standard hosted tiers as of
early 2026. Prices may have changed - check their websites for current
pricing.
Why is VecLabs so much cheaper?
Three reasons:
1. Decentralized storage economics - Shadow Drive storage costs are near the marginal cost of disk space with no corporate margin on top. Centralized providers charge 10-100x the raw storage cost.
2. No per-query fees - queries run against the in-memory HNSW index. There’s no compute charge per query - you’re not paying for cloud instances to process each request.
3. No platform tax - VecLabs is in alpha and not yet charging a platform fee. When we do introduce pricing, it will be significantly below incumbent providers.
Alpha pricing
During alpha, VecLabs charges $0 in platform fees. You pay only the direct infrastructure costs above (Shadow Drive + Solana fees).
When mainnet launches, we’ll introduce transparent pricing that remains well below Pinecone’s rates. Alpha users will receive preferential early-adopter pricing - join the waitlist at veclabs.xyz to lock in your rate.
Frequently asked questions
What happens to my data if I stop paying?
Your wallet controls your Shadow Drive storage. If you stop renewing storage, Shadow Drive nodes will eventually remove your data - the same as not renewing any cloud storage subscription. Your Solana on-chain records remain permanently.
Can I self-host to avoid all costs?
Yes. VecLabs is MIT licensed. You can run the HNSW core on your own infrastructure. The Solana program is deployed to devnet and eventually mainnet - you can interact with it directly. You’ll pay Solana fees but no Shadow Drive costs if you handle your own storage.
Do queries cost anything?
No. Queries run against the in-memory index. There are no per-query fees.