Memory that
outlives the
context window.

Why not Redis, SQLite, or a vector store?

You almost certainly run a database already. None of them was built to remember the way an agent needs to.

A vector store recalls by similarity but has no notion of an event, a fact that gets corrected, or context that should expire. A cache forgets on its own timer. A relational table stores rows but can't rank by meaning. Agent memory is all of these at once — so AegisDB makes each its own kind, with the lifecycle it actually needs.

what each one remembers

redis      → fast, but evicts on its own TTL
sqlite     → durable rows, no recall by meaning
pgvector   → similarity only — no events, TTL, or graph
aegisdb    → events, facts, working + a graph, each lifecycled

Stop paying to re-teach the model.

Every session normally re-pastes your stack, conventions, and past decisions into the prompt. AegisDB keeps that knowledge out of the window and feeds back only what's relevant, per prompt — so you pay for the work, not for re-establishing context.

• Recall, don't re-explain — your stack, conventions, and past decisions are already known, so you stop re-pasting them each session.
• A relevant slice, not a dump — each prompt pulls only the top-ranked memories — ranked by similarity × importance — instead of replaying entire transcripts; the model never sees the rest.
• Short sessions, full knowledge — start fresh instead of dragging a giant transcript whose every turn re-bills the whole context.
• Distilled & shared — capture stores salient outcomes, not raw logs, and a team reuses context established once.

Fast enough to be invisible.

Recall sits in the agent's inner loop, so latency is the whole game. AegisDB keeps vector search sub-millisecond as memories pile up — exact while a set is small, an HNSW graph past a threshold, so recall stays fast without losing accuracy.

These are AegisDB's own numbers, not a comparison — and you can reproduce them yourself: make bench (vector recall + latency) and make wire-bench (end-to-end over TCP).

make bench · make wire-bench

vector recall  → 0.14 ms @ 10k · 0.22 ms @ 100k
recall@10      → 1.00  (384-dim, HNSW)
ping           → 158k ops/s · p50 0.05 ms
read (get)     → p50 0.8 ms over the wire

measured on a 2023 laptop i7, loopback

Long-term memory for Claude Code.

An MCP server plus session hooks: relevant memories are recalled into context on every prompt, and the ones worth keeping are captured when the session ends.

• Automatic recall on each prompt, inside a strict time budget.
• Automatic capture at session end, filtered by salience.
• Per-project namespaces — isolated by default, shared by choice.
• No clone — uvx fetches the published package (aegisdb-mcp on PyPI) on demand.
• One-command setup — uvx --from aegisdb-mcp aegisdb-init writes the config below and the hooks; or install the /aegis-setup skill and let Claude do it.
• Fails open — if the backend is down, your session keeps working.

.mcp.json — scaffolded by aegisdb-init

{
  "mcpServers": {
    "memory": {
      "command": "uvx",
      "args": ["aegisdb-mcp"],
      "env": {
        "AEGIS_NAMESPACE": "my-project",
        "AEGIS_EMBEDDING_DIMENSIONS": "1024"
      }
    }
  }
}