Memory Hygiene Audit

AI Agent Memory Audit / Context Hygiene SaaS

Audit persistent agent memory before it rewrites the next run.

Memory Hygiene Audit scans CLAUDE.md, AGENTS.md, MCP memory notes, and coding-agent handoffs for stale facts, prompt-like poisoning, context bloat, unsafe retention, and missing provenance.

View pricing plans
Supermemory-style memory stacks Claude Code and Codex context MCP memory poisoning review
Memory hygiene scanner
64hygiene score
High riskHold shared memory promotion

Secrets, stale dates, and permanent incident language need review before agents can reuse this context.

Next evidenceUnlock source-aware report, cleanup diff, and agent-ready JSON.

Product workflow

Turn messy memory into an evidence-backed recall brief

Supermemory shows why persistent AI memory matters. Memory Hygiene Audit focuses on the operational gap around it: what memory should be kept, corrected, expired, blocked, or promoted before the next agent run depends on it.

Source-aware audit

Classify memory by trusted user input, repo fact, untrusted external text, MCP tool output, temporary incident note, or personal preference.

Context bloat score

Find duplicated instructions, oversized handoffs, dead commands, stale dates, and low-signal sections that waste attention.

Poisoning signals

Detect instruction hijacks, credential prompts, hidden external-action triggers, and false facts that should remain blocked.

Retention policy

Assign lifetimes, scopes, review owners, and deletion paths for memories before they become durable team context.

Agent-ready output

Export compact HTML, JSON, and llms-friendly summaries that agents and reviewers can cite.

Checkout-backed SaaS

Use direct NOWPayments hosted checkout for the highlighted Team Hygiene annual plan and public pricing pages.

Audit receipt

Every memory decision should explain its source, scope, and lifetime.

The scanner produces a reviewer-friendly receipt: what was kept, what was rewritten, what expires, what is blocked, and what an agent may cite later. The goal is usable continuity without invisible stale context.

Read the retention policy guide
Sourcerepo instructions, handoff note, MCP output
Riskstale fact, poisoning language, secret retention
Actionrewrite, expire, block, promote after review
OutputHTML report, JSON policy, compact agent brief

Workbench preview

A concrete console for memory cleanup, not a generic landing page.

Memory Hygiene Audit console showing risk score, retention policy, and cleanup actions

Pricing

Start with Team Hygiene annual, then scale policy coverage

Annual billing is selected by default and saves 50% compared with month-to-month billing. The Team Hygiene plan is highlighted because shared context cleanup is where repeat agent work starts paying back.

Team Hygiene annual is billed $1788 yearly.

Solo Audit

One operator

200 scans included for memory audit, context hygiene, and policy report workflows.

$49/mo

Billed $588 yearly

  • CLAUDE.md and AGENTS.md audits
  • Context bloat score
  • Poisoning signal preview
  • Cleanup receipt export
Default choice

Team Hygiene

Agent-heavy teams

2,000 scans included for memory audit, context hygiene, and policy report workflows.

$149/mo

Billed $1788 yearly

  • Everything in Solo Audit
  • Repo context diffing
  • MCP memory poisoning checks
  • Retention policy report
  • Team PDF and HTML evidence exports

Governance

Platform governance

10,000 scans plus API included for memory audit, context hygiene, and policy report workflows.

$449/mo

Billed $5388 yearly

  • Everything in Team Hygiene
  • Agent-ready JSON output
  • Private policy packs
  • Procurement review support
  • Priority support channel