Search vectors, traverse the graph, and scan recent history in a single call - one internally consistent answer, not three systems stitched together.
Agent memory & context platform
Every kind of agent memory - in one query.
Execution logs, vector recall, and reasoning graphs live in one store per agent - not a vector database bolted to a separate graph. So you commit a whole reasoning step atomically, query across all of it at once, and replay what your agent knew at any moment.
- Single region
- Multi-region geography
- Global
No credit card required · Cancel anytime
One store. Not a stack.
Most agent-memory tools bolt a vector database onto a separate graph. One store for all of it is what makes these four guarantees possible at once.
Commit a reasoning step's log, embeddings, and graph edges together. The instant it lands, every part is readable - no sync lag between stores.
Replay your agent's complete memory - logs, vectors, and graph together - exactly as it was at any past moment.
Every kind of memory sits behind a single tenant-isolation boundary. One thing to secure, one thing to audit.
Everything an autonomous agent needs
Relational, graph, and vector in a single ACID store that spans the globe.
Query across all memory at once
Semantic recall, graph reasoning, and recent history in one request - fused into a single ranked, consistent result instead of app-side glue across separate stores.
Atomic writes, zero sync lag
Write a step's log, embeddings, and graph edges in one transaction. Read-your-writes across every memory type - no eventual-consistency gaps between systems.
Time-travel any moment
Read an agent's full memory exactly as it was at any past instant - logs, vectors, and graph together - for debugging, audit, and reproducibility.
Provable, atomic erasure
Delete an agent and everything it remembered vanishes in a single operation - returning a receipt you can hand to a compliance review.
Pinned where your agents run
Each agent's memory is pinned to its home region and served from the nearest data plane, so reads and writes stay ultra-low-latency - in a single region, a whole geography, or the entire globe.
API & SDK first
Spawn agents, write memory, and run retrieval through a clean API and typed SDKs - drop Jennah into your agent framework in an afternoon.
Simple, scalable pricing
Start free for 14 days, no credit card required. Every plan includes storage and a monthly query allowance - add more storage anytime at a flat per-GB rate.
Pro
$99/mo
25 GB storage included · then $0.50/GB
1M queries / mo included
For teams launching agents in a single region.
- Single region of your choice
- Vector search & in-database inference
- Unlimited projects
- Up to 100 active agents
- Priority support
No credit card required
Scale
$299/mo
60 GB storage included · then $0.50/GB
5M queries / mo included
For teams outgrowing Pro in a single region.
- Everything in Pro
- Up to 500 active agents
- Higher throughput limits
- Priority support
Team
$999/mo
150 GB storage included · then $0.75/GB
20M queries / mo included
For teams that need regional redundancy.
- Everything in Scale
- Multi-region within one geography (US, EU, or Asia)
- Higher agent & throughput limits
- SSO
- Priority support with faster response SLAs
Enterprise
Custom
Custom storage · volume pricing
Custom query volume
For global-scale, regulated deployments.
- Everything in Team
- Global, cross-geography multi-region
- Dedicated regional instances
- Custom data residency
- SSO & audit logging
- SLA & dedicated support
A query is a vector search or graph traversal; traversals and cross-modal queries count by depth. Generous fair-use limits apply during early access. Prices cover compute and storage. AI inference - embeddings and model calls - is billed separately based on usage.
Talk to us
Start your free trial or ask us anything - we usually reply within one business day.
Prefer email? Reach us at info@nightblue.io.