What will memory cost?
Enter your workload to see what each managed memory layer or vector store would charge on its published plans and rates, with the arithmetic and the pricing page behind every number. When your usage is beyond a vendor's largest published plan, you'll see "Contact sales"; when a vendor publishes no rates, "Not published". Never a guess. Self-hosting the open-source versions costs only your infrastructure, which isn't estimated here.
-
- ~9,000,000 tokens a month (600 bytes per message ≈ 150 tokens)
- Standard: $1 per 1M tokens
One workspace assumed; a published per-token rate is applied to all tokens.
Pricing page · checked 2026-10-27
-
- ~9,000,000 tokens written × $2/1M + 60,000 searches × $0/1M = $18 of usage
- Pay as you go (card on file): $0 incl. $5 credit, + $13 extra credits
Pricing page · checked 2026-10-27
-
- ~9,000,000 tokens written × $5/1M + 60,000 searches × $5/1M = $45.30 of usage
- Pro: $19 incl. $20 credit, + $25.30 extra credits
Pricing page · checked 2026-10-27
-
- 1,000 agents (1 per user)
- Developer: $0.1 per active agent
Plus LLM and tool-execution usage, which the vendor bills separately and this doesn't estimate.
Pricing page · checked 2026-10-27
-
- 60,000 episodes × 2 credits (600 bytes ÷ 350 per credit) = 120,000 credits
- Flex: $125 incl. 50,000 credits + 7 top-ups × $25
Pricing page · checked 2026-10-27
-
- 60,000 memory writes and 60,000 searches a month
Usage is beyond the largest published plan; the vendor quotes this privately.
Pricing page · checked 2026-10-27
Estimates assume one memory write per message and about 4 bytes per token. They cover the memory layer's own fees only; LLM calls your app makes are extra, and so are LLM calls the layer bills separately. Prices change, so confirm on the vendor's page before you commit. Vector-store estimates assume one namespace / collection and 4 bytes per dimension. Missing a vendor, or found a price that changed? Edit the memory or vector-store pricing file. To pick a layer on features first, use the chooser.