CockroachDB × AWS — agentic memory

Agent memory that survives the outage.

BlackBox is agentic-memory infrastructure with two properties most agent memories lack: it survives a full region failure, and it audits what it lets itself remember — every learned fix is gated, deduplicated, and confidence-scored before it can shape recall. We prove both with an incident-response agent working a live outage on CockroachDB.

memories on record3 regionsvector search
memory · regional by row recording
aws-us-east-1
0
memories pinned
aws-eu-west-1
0
memories pinned
aws-ap-south-1
0
memories pinned
survive region failuredemo · 0 total
Built on
CockroachDB CloudDistributed Vector IndexingManaged MCP ServerAWS Bedrock — ClaudeTitan Text Embeddings v2Next.js · TypeScript
Capabilities

Agent memory is a distributed-systems problem.

An agent's memory is only trustworthy if it is available during a failure, consistent across regions, compliant with where data may live — and honest about what it lets itself remember. Bolt a vector store onto a cache onto a state store and none of that holds when a region goes dark. BlackBox runs Claude via Amazon Bedrock over one system of record — and everything below reads from the production cluster.

Survive

Kill a region. The memory keeps answering.

us-east-1eu-west-1ap-south-1

Every memory replicates across three AWS regions under SURVIVE REGION FAILURE. Lose one mid-incident and recall keeps serving from the survivors — including rows homed in the dead region. The counts above are live from the production cluster.

last validated drill · 9-node local rigprimary region killed · top-5 recall identical · 136 ms
Watch the failure drill →
Recall

Institutional memory, in milliseconds

vector search, measured live
memories on record

The vector index is distributed like everything else — search runs next to the data in each region, so recall survives whatever the data survives.

Evidence

Every answer cites its memory

recall captured 2026-07-28, production cluster · lower = closer · ×N = recurrences
[1]checkout-api p99 latency spike to 5s from connection pool exhaustionap-south-1 · 0.74 · ×72
[2]checkout-api thread pool starvation under burst trafficap-south-1 · 0.80 · ×48
[3]Learned runbook: checkout-api connection pool exhaustioneu-west-1 · 0.78
Run your own recall in the console →
Compliant

Pinned to its home region

ap-south-1
eu-west-1
us-east-1

REGIONAL BY ROW domiciles each memory where the incident happened. An EU incident’s data physically stays in the EU — no second database.

Learn

Runbooks that compound

Resolving an incident distils what worked into a new runbook, written back to the same memory. The next similar incident recalls it first — the agent gets sharper with every failure it survives.

How the agent works →
Introspect

Reads its own cluster

inspect_cluster → select_query, via the Managed MCP Server

Through CockroachDB’s Managed MCP Server the agent queries the very database it runs on — mid-drill, it diagnosed its own region outage and reported memory still serving.

Why CockroachDB, in depth →

Watch an agent remember through a region failure.