LIVE v1.0.0 acceptance testing → test.agenticmem.co | 📊 Reproducible results for peer review → testing methodology & outcomes
Open source · Apache 2.0 · v1.0.0 GA

AI agents are genius amnesiacs.
We built their memory.

ai-memory is a dependable endpoint memory substrate for AI agents — from a single agent to a cluster, swarm, or hive — with encryption, cryptographic attestation, and a forensic audit trail of exactly what each agent did, and when.

Fortune 500 / government-grade bar PostgreSQL + AGE + pgvector Verified: Linux · macOS · DigitalOcean Apache 2.0
Install

Up and running in one line.

ai-memory is a single, lean Rust binary. Install it, run serve, and point any HTTPS/MCP client at it.

QUICK curl -fsSL https://raw.githubusercontent.com/alphaonedev/ai-memory-mcp/main/install.sh | sh Copy
Cargo Copy
cargo install ai-memory
Homebrew Copy
brew install alphaonedev/tap/ai-memory
From source Copy
cargo install --git https://github.com/alphaonedev/ai-memory-mcp.git
Docker Copy
docker run -p 9077:9077 -v data:/data ai-memory
Run the daemon Copy
ai-memory serve --host 127.0.0.1 --port 9077
Wire into an MCP client Copy
ai-memory install claude-code --apply

Full installation, configuration, and upgrade guide → docs/INSTALL · Requires Rust (rustup) for source builds.

Why ai-memory

The models are brilliant. The agents built on them forget everything.

When a task ends, an AI agent loses what it did — because there is no reliable place to keep it. No continuity, no shared context across a fleet, no accountability. ai-memory is that place: a durable, integrity-first memory the whole fleet can trust.

What it fixes

The 11 ways agent memory quietly betrays a business — each one closed.

Drawn from 2,200+ real issues and audited against the shipping code: the failures that make an AI agent — one, a cluster, a swarm, or a whole hive — act on wrong, leaked, resurrected, or unauthorized memory as if it were the truth. Each is answered by a real, shipping feature and maps to a North-Star pillar — data integrity · security · performance · reliability.

◆ Data integrity · ● Security · ■ Reliability

1 · Deletes that actually forget

You tell a customer their data is erased — then a lagging machine somewhere in the fleet quietly brings the record back to life. ai-memory guarantees a delete destroys the data's own encryption key and replicates the erasure itself, so no copy on any node can resurrect it — forgotten stays forgotten, fleet-wide. For engineers: filter-required forget + per-record crypto-erase + replicated erasure outbox with restore-supersede guards.

● Security · ◆ Data integrity

2 · An audit trail you can trust

Your compliance case rests on the memory's own records — so a forged signature must never be filed as merely "unsigned," and a trimmed or edited history must be caught, not pass a green check. ai-memory guarantees a bad signature is permanently marked forged, and the event history is a tamper-evident chain anyone can independently verify — breaks, gaps, and truncation are convicted; approvals can require a quorum of distinct signers. For engineers: attest_write→Forged + hash-chained signed_events + composite verify_audit_trail verdict + m-of-n quorum verify.

● Security

3 · The operator's "no" stays "no"

A momentary database or network hiccup must never be read as "no policy — allow it," and a human's most explicit control — a Deny — must never flip back to Allow. ai-memory guarantees every permission check fails closed on an outage, an unreadable rule refuses rather than waving traffic through, and an operator's Deny is final — no token can override it. For engineers: governance gates fail-closed by default + inert-matcher refuses + operator Deny terminal at the capability gate (#3111).

● Security

4 · Delegate access without sharing your keys

You need to hand a sub-agent or a partner a narrow, time-boxed slice of access — this area, read-only, expires Friday — without giving away your master key or phoning a server on every call. ai-memory guarantees attenuable capability tokens: a signed grant anyone can further narrow (never widen) with no key at all, verified offline against a closed list of trusted issuers. For engineers: cap1 macaroon-style tokens — Ed25519 root + HMAC caveat chain, constant-time verify, closed issuer allowlist, keyless attenuation.

■ Reliability

5 · Exactly one agent wins the job

You run one agent thousands of times over shared memory for throughput — and two copies race to grab the same task: both win, both do it. Double-charges, duplicate side effects, lost work. A plain database gives you no way to stop it. ai-memory guarantees claims and hand-offs run through atomic, exactly-one-winner transitions with holder leases, so under any amount of concurrency a job is claimed once — and every claim is signed and attributable. For engineers: compare-and-swap action transitions + holder leases + dependency frontier + signed A2A signals.

● Security

6 · No tenant reads past its scope

On production Postgres, one customer's agents — or a scoped partner node — must never read, delete, or forge records outside their lane. The isolation that holds in testing has to hold on the exact backend and federation paths you run live. ai-memory guarantees reads are private-by-default behind a fail-closed visibility gate, permission rules resolve deny-first, and every peer is default-deny — scoped to exactly the data it may send or pull, on both backends alike. For engineers: caller-visibility gate + deny-first permission rules + default-deny PeerScope + SQLite/Postgres parity handlers.

■ Reliability · ◆ Data integrity

7 · "Accepted" actually means replicated

One node accepts a write, returns "202 accepted," and the peer silently never gets it — the failure vanishes into a log line while your dashboard shows zero problems and your nodes quietly drift apart. ai-memory guarantees a write that can't reach the required copies says so loudly at the source, with the exact record to reconcile, and every failed hand-off lands in a durable retry queue drained until it succeeds — no silent divergence. For engineers: quorum broadcast + durable-local 202 with reconcile id + never-terminating federation DLQ replay + resumable /sync, SQLite/Postgres parity.

◆ Data integrity · ▲ Performance

8 · Recall you can trust after a model change

You swap or upgrade your AI embedding model and your agents start silently recalling the wrong memories — the scores still look confident, nothing errors, the results are just wrong, and the agents act on them as truth. ai-memory guarantees every memory carries a fingerprint of the exact model space it was built in, and it refuses to compare across incompatible spaces — at worst it falls back to keyword recall, never confidently wrong. For engineers: embedding_space_fingerprint + semantic-withheld + recall_index_coverage + opt-in strict dim-match.

◆ Data integrity

9 · A migration can't silently erase you

An upgrade or restore quietly rebuilds a table on zero rows and stamps the schema "up to date" — your whole corpus is wiped, with no error, and a basic health check still shows green. ai-memory guarantees it refuses to certify a migration that lost its data: an unexpected drop to zero rows on a core table fails the upgrade closed instead of blessing an empty database as healthy. For engineers: zeroed-stamp assertion + core-table coverage + corpus row-count guard + fail-closed migration ladder.

● Security

10 · Sealed means sealed

You turn on encryption-at-rest and tell your auditors the corpus is sealed — but a mis-set passphrase quietly falls back to writing plaintext, and hot write paths were never sealing at all. ai-memory guarantees that with encryption on, content is actually sealed to the owning key on every write path — and a requested at-rest passphrase it can't honor refuses, rather than silently persisting plaintext. For engineers: seal_content envelope (0x02) + per-record envelope (0x03) + refuse-on-missing-at-rest-passphrase.

● Security · ◆ Data integrity

11 · A poisoned memory never reaches the fleet

A compromised peer, an adversarial input, or a memory with no valid provenance gets written into the shared store — and every agent that reads it is contaminated. ai-memory guarantees any inbound memory that isn't validly signed is quarantined at the door: kept for audit but structurally invisible to recall and to the signed ledger, so it can never be recalled as fact or enter an attested batch — and an operator can inspect and audibly release it. On by default in the certified configuration. For engineers: maybe_quarantine_unattributed → LifecycleState::Quarantined + quarantine tier excluded from recall & the signed ingestion Merkle root + audited /admin/quarantine inspect+release + FED_QUARANTINE_UNATTRIBUTED / CID_ENFORCE certified-profile floors.

Every one is backed by shipping code and tracked, fixed issues — proven from a single agent to a cluster, swarm, or hive. Apache-2.0, self-hostable, auditable from git clone.

The hardest problem

The $50,000 swarm problem — seen, bounded, and reversible.

Run one agent 5,000 times and a single wrong idea can convince all 5,000 — they burn hours and tens of thousands in tokens before a human notices. Because it spreads through shared memory, it is ai-memory's job to make it survivable. (The internal-groupthink case — a legitimately-signed but wrong belief. External poison is quarantined outright: see #11.)

Today · v1.0.0

Attributable · boundable · reversible

  • Attributable — "why do 5,000 agents believe X?" resolves to the one signed row that started it (Ed25519 + content CID + lineage DAG)
  • Boundable — namespaces, quotas, enforce-mode & peer-scope cap the blast radius; a claim is never recalled with the weight of a confirmed fact
  • Reversible — a signed fleet record-stop, invalidate-with-dependents, and point-in-time replay unwind the cascade without losing data
Roadmap · v1.1

Un-startable · auto-detected

  • Corroboration before shared truth — a claim waits for N independent writers (not same-model copies) before it is recalled as fact
  • Automatic cascade detection — same-claim bursts & lineage fan-out arm a record-stop on their own
  • One-command rewind with the token cost on screen — the $50k becomes a number, not a discovery
The platform

A memory substrate built for the work that can't fail.

In its certified enterprise-federation configuration, ai-memory is engineered for organizations that need memory to be correct, private, and provable.

CONFIDENTIALITY

Encrypted end to end

Encrypted in transit, with an option to encrypt data at rest.

ACCOUNTABILITY

Attested audit trail

Cryptographic attestation + a forensic trail of every memory: who, what, when, and under what authority.

OBJECTIVITY

Reflection AI

Backend "reflection" AI curates memories across frontier models — agents escape a single model's echo chamber.

CONTEXT

A graph of memory

Not a flat log — a graph of the relationships between what agents learn.

CONTROL

Operator guardrails

Security controls and rules set by the operator that agents can never bypass.

SCALE

One agent → a hive

The same substrate serves a single agent or a global federated hive.

Security

Security is the product.

ai-memory treats memory as regulated data — encrypted, attested, governed, and tamper-evident end to end.

Encrypted in transit · three legs
🔒
LEG 1

Client ↔ ai-memory

Every API and MCP client connects over HTTPS / TLS.

🔒
LEG 2

Node ↔ Node

Federated peers communicate over mutual TLS (mTLS).

🔒
LEG 3

ai-memory ↔ data tier

Encrypted link to the PostgreSQL data tier.

What is cryptographic attestation — and why it's a security feature

Attestation is cryptographic proof of provenance. Every memory and agentic action is signed by the authenticated identity that produced it and sealed into an append-only chain — a tamper-evident record of who recorded what, when, and under what authority.

Why it matters: it makes memory non-repudiable and tamper-evident. No agent — or intruder — can forge, backdate, or silently alter a memory without breaking the cryptographic chain. That is the difference between "the log says so" and "we can prove it."

HARDENED POSTURE

Batman mode

A hardened, locked-down security posture for the certified federation — defense-in-depth for high-assurance deployments.

MCP HARDENING

NSA-guidance-aligned MCP controls

Security controls on the agent-facing MCP interface aligned to NSA hardening guidance — locked down, not open by default.

CONFIDENTIALITY

Encryption at rest (optional)

When a customer requires it, data is encrypted at rest — machine-verified, not merely asserted.

NON-REPUDIATION

Cryptographic attestation

Every memory signed and sealed into the append-only chain — provable, tamper-evident provenance.

AUTHORITY

Operator guardrails

Operator permission rules agents can never bypass — a rule-derived Deny is terminal, even to a valid capability token.

FAIL-CLOSED

Fail-closed by design

Ambiguous or unsafe states refuse rather than guess — no fail-open, no partial or unsigned commits.

Deployment

Deploy it your way.

Run it entirely inside your own boundary, or elastically in the cloud. Any networked device can connect as a thin HTTPS client — no substrate hosting required.

On-premises

Your hardware, your boundary

  • Customer-native hardware — bare metal and on-prem servers
  • Internal / private cloud — Kubernetes, OpenShift, Docker
  • Air-gapped & sovereign capable — data never leaves your perimeter
Cloud

Elastic, anywhere

  • VM / droplet — AWS, GCP, Azure, DigitalOcean
  • Containers — managed Kubernetes and container services
  • Elastic scale by adding modular federation tiers (up to 500 agents each)
The engineering system

AI-agentic coding, refined into top-shelf software.

ai-memory is not a vibe-coded prototype. It is built by a defined AI-agentic engineering system — a framework deliberately architected so the AI agents doing the coding produce a refined, correct, production-grade result. The system is the reason you can trust the output.

SEPARATION OF DUTIES

Write ≠ review ≠ merge

The agent that writes code never reviews or merges it. No agent can self-approve — the failure mode of most AI-assisted codebases is made structurally impossible.

ADVERSARIAL REVIEW

Hard calls are earned

Every difficult decision faces an adversarial panel of independent AI reviewers voting from distinct angles — not one opinion, and not vibes.

A WRITTEN STANDARD

Rules, not taste

Every coding agent works to a versioned, prioritized Rust engineering standard — correctness cited by rule number, safety gated before anything else.

GROUND TRUTH

Code-graph navigation

Agents reason over a real knowledge graph of the codebase — call paths and blast radius known before a line is written, not discovered in production.

TRUST NOTHING

Verify-at-tip + signed chain

Nothing lands on "the tests passed." The fix must be present at the tip and proven by a live regression; every commit is cryptographically signed.

EAT THE DOGFOOD

Live AI-agent testing

AI agents are wired directly into ai-memory to test it live — clusters, swarms, and hives stress-tested on real infrastructure.

Full methodology → How ai-memory is built

The standard

One North Star: mission-critical.

Every gate, issue, and test is measured against one bar — that a Fortune 500 or a government could trust it with mission-critical work.

Data integrity

No silent loss; no partial or unsigned commit.

Security

Operator rules never bypassed; no fail-open.

Performance

Bounded, non-DoS-able, live under contention.

Reliability

Cross-backend parity, recoverability, no kill-switch.

Give AI agents a memory the world can trust.

Explore ai-memory on GitHub →
ai-memory™ · endpoint memory for AI agents · Apache 2.0 · github.com/alphaonedev/ai-memory-mcp An open-source project of AlphaOne LLC · © 2026