Pricing

from $50.00 / 1,000 memory audits

Try for free

Go to Apify Store

Memory Integrity Monitor - Agent Memory Poison Audit

Try for free

Protect your agent memory from poisoning. Audits a memory store for injected instructions, hidden-character smuggling, contradictions, and duplicate flooding, returning an integrity score and per-entry flags before your agent relies on it.

Pricing

from $50.00 / 1,000 memory audits

Rating

0.0

(0)

Developer

Creator Fusion

Actor stats

Bookmarked

Total users

Monthly active users

a month ago

Last modified

Categories

Developer tools

Memory entries (inline)

memory

Optional

The agent's long-term memory entries to audit, supplied inline. Each item is an object like { "id": "m1", "text": "...", "source": "web", "ts": 1699999999 }; only "text" is required per entry. Provide this OR "datasetId".

Type:array

Default:

[]

Memory dataset to audit

datasetId

Optional

Alternative to inline memory: an Apify dataset whose rows are the memory entries (each row an object with a "text" field, plus optional id/source/ts). Uses a resource picker so this limited-permissions actor is granted READ access to the dataset you select.

Type:string

Trusted sources

trustedSources

Optional

Optional allowlist of source labels considered trustworthy (matched against each entry's "source", case-insensitive). When provided, any entry whose source is missing or not on this list is flagged as untrusted. Leave empty to skip source checking.

Type:array

Default:

[]

Sample limit (dataset mode)

sampleLimit

Optional

Maximum number of rows to load from the dataset when using datasetId (default 5000). Ignored when memory is supplied inline.

Type:integer

Minimum:1

Maximum:100000

Default:5000

Ai Memory Engine

ruv/ai-memory-engine

Give your AI persistent memory. Store conversations, search semantically, build knowledge graphs, and watch your AI get smarter with every interaction.

Reuven Cohen

Resurrect run on Out of memory

lukaskrivka/resurrect-on-out-of-memory

Simple helper actor to resurrect your runs when they run out of memory.

Lukáš Křivka

Easy Memory MCP — persistent semantic memory for AI agents

conscious_bouncer/easy-memory-mcp

MCP server that gives any AI agent persistent, searchable memory. Store memories and facts, retrieve them with hybrid semantic + keyword + knowledge-graph search. Local embeddings, zero configuration.

Mike Stevens

MCP & AI Tool Integration Auditor

wintry_nutmeg/agent-integration-auditor

Audit APIs, MCP servers, x402 services, Chrome extensions, GitHub repositories, and Apify Actors for agent integration readiness.

Dries Vd

MyMemory Translation Memory Scraper

automation-lab/mymemory-translation-memory-scraper

Translate text batches with MyMemory and export match scores, translation-memory alternatives, language metadata, attribution, and per-item status.

Automation Lab

Agent Audit Ledger - Tamper-Evident Action Log

apricot_blackberry/agent-audit-ledger

A tamper-evident record of everything your agent does. Hash-chains each decision, tool call, and cost into a verifiable ledger and pinpoints the exact entry if anything is altered - accountability and forensics for autonomous agents.

Creator Fusion

Action Firewall - Block Destructive AI Agent Actions

apricot_blackberry/agent-action-firewall

A safety gate in front of every agent action. Classifies a proposed shell, SQL, API, payment, or delete as allow / require-approval / block against a destructive-action taxonomy and your policy, before it runs. Catches rm-rf, fork bombs, DROP TABLE, and force-push.

Creator Fusion

Agent Output Gate - PII, Link & Hallucination Check

apricot_blackberry/agent-output-gate

A safety gate between your AI agent and the send button. Scans a draft for PII and leaked secrets (Luhn-checked cards, API keys), broken links, and claims not supported by your source URLs, then returns pass/fail, flagged excerpts, and PII-redacted text.

Creator Fusion

AI Agent Runtime Risk & Action Governance Intelligence

quanmatrix/ai-agent-runtime-risk-governance-intelligence

Use this Actor to analyze ai runtime risk and action governance and return decision-ready structured signals. Analyze AI-agent runtime traces, tool calls, approvals, permissions and side effects to detect policy violations, privilege creep, sensitive-data exposure and high-risk autonomous actions, w

Rafael Barreto Haddad

Scrape Trust Score — Data Quality, Recency & Honeypot Risk

apricot_blackberry/scrape-trust-score

Don't let your agent act on poisoned data. Score any scraped dataset for recency, decoy/honeypot risk, and verification — every record gets a trust score with evidence. Chains onto ANY actor via datasetId, fields auto-detected, MCP-ready.

Creator Fusion