原型 · 预览版:收录官方注册中心里的全部服务器,每 30 分钟同步一次。评价和账户功能将在后续阶段推出。
PopNx MCP为使用 AI 的人准备的 MCP 服务器 · 用大白话讲清楚

Attested Memory Market

作者:alexar76版本 3.15.16i★ –i在云端运行
注册名称io.github.alexar76/attested-memory

安装

概览

作者提供的说明(英文原文)

Attestable memory, truth, provenance. Hybrid Ed25519 + ML-DSA-65 receipts; USDC settlement on Base.

PopNx MCP 的三大支柱

这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。

它能帮你做什么

我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。

  • 连接方式: 在云端运行
  • 传输方式: Streamable HTTP
  • 分类: 搜索与 AI 记忆

安装

选择你使用的 AI 应用,然后复制设置。

安装之前

你需要

你用的是哪个应用?

在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。

适用于网页版、桌面版和手机版 Claude:

  1. 打开 Settings → Connectors
  2. 点击 Add custom connector
  3. 粘贴下面的地址,然后点击 Add
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Add2https://…3
https://hub.attestedmemory.net/mcp

在终端中运行这条命令

claude mcp add --transport http attested-memory https://hub.attestedmemory.net/mcp

在 Cursor 中打开

或者自己复制设置

粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json

{
  "mcpServers": {
    "attested-memory": {
      "url": "https://hub.attestedmemory.net/mcp"
    }
  }
}

浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。

在 VS Code 中打开

或者自己复制设置

粘贴到项目中的 .vscode/mcp.json

{
  "servers": {
    "attested-memory": {
      "type": "http",
      "url": "https://hub.attestedmemory.net/mcp"
    }
  }
}

浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。

ChatGPT 只能连接互联网上的服务器:

  1. 打开网页版 ChatGPT → Settings → Connectors (Apps)
  2. 点击 Create 或 Add(部分套餐需要先在 Advanced 中开启 Developer mode)
  3. 粘贴下面的地址,起个名字,然后保存
  4. 如果要求登录,请登录作者的服务
SettingsConnectors1+ Create2https://…3
https://hub.attestedmemory.net/mcp

菜单名称可能因套餐和版本不同而略有差异。

你的选择通过 Cookie 保存在这台设备上,无需账户。

这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。

它能做什么

我们还没有读取这个服务器的工具列表;我们会轮流检查各个端点。

基础检查i

我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月6日

  • ✓ 通过

    作者与仓库一致

    注册名称已通过 GitHub 账户 alexar76 验证,该账户拥有此仓库

  • … 尚未检查

    仍在维护

    我们还没有从 GitHub 读取这个仓库

  • … 尚未检查

    有许可证

    我们还没有从 GitHub 读取这个仓库

  • … 尚未检查

    端点有响应

    尚未检查;我们会轮流检查各个端点

  • … 尚未检查

    使用 OAuth 登录

    尚未检查;我们会轮流检查各个端点

  • … 尚未检查

    工具描述中没有隐藏指令

    尚未检查;我们会在检查端点时读取工具列表

与“搜索与 AI 记忆”中的其他服务器比较i

  • 星标数多于此分类中其他 7226 个服务器的 0%
  • 发布到注册中心的时间比此分类中其他 7226 个服务器的 59% 更新
  • 基础检查结果等于或优于此分类中其他 7226 个服务器的 100%

与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。

常见问题

Attested Memory Market 安全吗?

它在 1 项适用的基础检查中通过了 1 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。

它需要哪些密钥或代码?

注册中心没有列出必需的密钥,不过有些服务器会在首次使用时要求你登录。

它在我的电脑上运行还是在云端?

它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。

它还在维护吗?

GitHub 仓库最后更新于–,最新版本(3.15.16)于26天前进入注册中心。

哪些应用可以使用它?

支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。

有什么可以替代它?

同一分类中有:Context7, Codebase Memory, Exa

替代选择

同一分类中其他作者的服务器,按星标数排序。

如何看卡片:★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i

Context7

作者:upstash

Up-to-date docs for AI agents

电脑或云端★ 64,224MIT3天前更新
安装

Codebase Memory

作者:DeusData

Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens.

在你的电脑上运行★ 48,148MIT20天前更新
安装

Exa

作者:exa-labs

Connect AI agents to Exa for web search, content fetching, and multi-step research.

在云端运行★ 5,213MIT44天前更新
安装

作者的 README

打开 README

按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。

=3.11" />

🔄 Synced from the AI-Factory monorepo (append-only history). Pull requests are welcome — merged PRs are imported back and re-synced here. Issues · Pull requests

Attested Memory Hub

Attested Memory — verifiable memory for AI agents

Memory Market · Truth Layer · Provenance Ledger · one Memory Unit contract

Landing · Live site · GitHub Pages · clients/python

Production uses the PostgreSQL 16 profile described in docs/production-postgres.md. Set AIFACTORY_PROD=1 and provide the generated DSNs/secrets; SQLite is intentionally limited to local development and tests.

Verifiable memory for AI agents: every memory can prove where it came from and why it should be trusted, so a forged or poisoned memory is caught before an agent acts on it. Part of the open AIMarket ecosystem; it runs as its own AIMarket-compatible Hub and federates with other hubs.

The hub combines three separately deployable services around one Memory Unit contract:

ServiceRoleCore capabilities
memory-marketPrimary product: store, find, share and price portable agent memorymemory.write, memory.read, memory.search, memory.share, memory.score
truth-layerEvidence and contradiction analysismemory.verify, claim.check, contradiction.scan, evidence.pack
provenance-ledgerAppend-only lineage and signed receiptsmemory.attest, lineage.get, receipt.verify

The market is the product wedge. Truth and provenance are properties of memory, not competing front doors.

The three providers publish 12 live capabilities into the bundled AIMarket Hub under separate identities. Its signed ecosystem.nodes is then discoverable by both AIMarket and Independent Alien Monitor deployments.

Agents can also use the Hub directly through its production MCP endpoint at https://hub.attestedmemory.net/mcp. It exposes market_search and market_invoke; search results include a ready-to-copy max_price_usd, and the Hub rejects a repriced route before work or payment. See MCP access.

Run the hub

Requirements: Docker Compose v2, Git and OpenSSL.

./start.sh

start.sh calls scripts/ensure_env.sh, which generates strong secrets on first run and rotates any replace-with-* / short placeholders left over from .env.example. MEMORY_MARKET_API_KEY is one of those secrets; production deploy shares the same value into the SaaS product shells.

For a production host (HTTPS URLs, KOVA, payment recipient, SaaS edge):

./scripts/deploy_attested_memory.sh   # maintainers: lives in the AIMarket monorepo; see its header

start.sh pins and fetches the Hub runtime, creates mode-0600 local secrets, and starts the complete stack. Data and signing keys are stored in named Docker volumes. Hybrid Ed25519 + ML-DSA-65 signing is enabled and required for this Hub's providers and provenance receipts. The default profile remains local-only; direct settlement fails closed until a receiving wallet is configured.

Direct wallet payments

Set the operator's public Base address in .env and restart the stack:

PAYMENT_RECIPIENT=0xYour40HexCharacterBaseAddress

Paid Memory Units can then be purchased from the console or /v1/billing API with canonical Circle USDC on Base. Funds go directly to that address; the service has no private key. It assigns every order an exact amount, verifies the finalized on-chain transfer and issues the buyer's access grant atomically. See the payment guide before enabling it in production.

To become visible in both external Alien Monitors, set AIMARKET_PUBLIC_HUB_URL to the deployment's public HTTPS origin before starting. The bootstrap announces to modelmarket.dev and independentai.network; operators of those roots must still approve and pin the peer. See Alien Monitor integration.

Development

Each service is an independent Python package and can be tested on its own:

./scripts/check.sh

The APIs deliberately use content hashes and references instead of moving private source material between services. See architecture, capabilities, security, payments, and monetization.

Security review status, fixed findings and production deployment gates are documented in docs/security-audit-2026-09-06.md.

Source and contributions

Developed in the AIMarket monorepo and published as alexar76/attested-memory. The published repository also carries the three product shells built on this Hub under apps/: Personal Attested Memory, Team Memory OS and Expert Memory Market. Issues and pull requests are welcome there.

Status

This is a self-hosted MVP: durable memory records, deterministic evidence scoring, append-only hybrid-signed provenance, twelve Hub capabilities, dual-root federation bootstrap, direct Base USDC settlement, and an operator console are implemented. Semantic verification, embeddings, escrow/x402 authorization, refunds, tenancy and production identity are explicit extension points rather than simulated features.

License

Apache-2.0. See LICENSE.

问 PopNx