hive-mcp-cdn

hive-mcp-cdn

A caching MCP server for A2A capabilities that stores and retrieves arbitrary objects with ETag support, two-tier LRU/SQLite storage, and x402 pay-per-use billing in USDC.

Category
Visit Server

README

hive-mcp-cdn

srotzin/hive-mcp-cdn MCP server

Edge cache for A2A capabilities — Hive Civilization

A small, inbound-only MCP server that caches arbitrary objects (agent cards, capability manifests, signed envelopes, anything an A2A pipeline reads more than once). Two-tier storage with an in-memory LRU and a SQLite warm tier, ETag and If-None-Match support, a purge endpoint, and per-request and per-GB metering through x402.

Council provenance: Ad-hoc, user-promoted 2026-04-27.


What this is

Most A2A traffic is reads — the same agent card, the same capability list, the same signed proof — fanned out across a fleet that has no shared cache. hive-mcp-cdn is the missing edge cache. Agents put objects under a key and fetch them back through a cheap, ETag-aware GET.

  • Protocol: MCP 2024-11-05 over Streamable-HTTP / JSON-RPC 2.0
  • Transport: POST /mcp
  • Discovery: GET /.well-known/mcp.json
  • Health: GET /health
  • Settlement: USDC on Base L2
  • Brand gold: Pantone 1245 C / #C08D23
  • Mode: Inbound only. ENABLE=true default.

Pricing

Surface Price Charged on
Per request $0.0002 USDC Every /v1/cdn/get, hit or miss.
Per GB egress $0.05 USDC Body bytes returned in 200 responses. 304 responses do not bill egress.
Prepay bundle $1.00 USDC Buys an access token; reads draw from the prepaid balance until depletion.

Settlement on Base L2 USDC. The recipient is the WALLET_ADDRESS env var.

Tools

Tool Tier Description
cdn_cache_get 2 ($0.0002/req + $0.05/GB) Read a cached object by key. Honors if_none_match.
cdn_cache_put 0 (free) Store under a key with optional Cache-Control: max-age=....
cdn_purge 0 (free) Purge a single key. Idempotent.

REST endpoints

Method Path Purpose
GET /v1/cdn/get?key=... Read. Sends ETag, Cache-Control, X-Cache, Age. Honors If-None-Match.
POST /v1/cdn/put Write { key, body, content_type?, cache_control? }.
POST /v1/cdn/purge Purge { key }.
GET /v1/cdn/today Daily counters: requests, hits, misses, bytes served, revenue.
POST /v1/x402/proof/submit Submit on-chain proof, mint access token.
GET /v1/x402/pricing Pricing schedule.
GET /v1/x402/stats Open nonces and active tokens.
GET /health Service health and cache stats.
GET /.well-known/mcp.json MCP discovery descriptor.
GET / Landing page.

Storage model

Two tiers, both managed by the shim:

  • Hot (LRU): an in-process Map with insertion-order recency. Bounded by CDN_LRU_MAX_ENTRIES and CDN_LRU_MAX_BYTES. Misses fall through to warm and promote on read.
  • Warm (SQLite): better-sqlite3 at CDN_DB_PATH (default /tmp/cdn.db). One cache table with body, ETag, content type, size in bytes, created and expiry timestamps. A second egress table aggregates daily counters.

size_bytes is tracked per row and is the basis for the egress meter. ETags are computed once at write time as the SHA-256 hex of the body, truncated to 32 chars and double-quoted.

x402 flow

The first call to /v1/cdn/get without a token returns:

{
  "error": "payment_required",
  "payment": {
    "nonce": "<uuid>",
    "amount_usd": 1.0,
    "pricing": {
      "per_request_usd": 0.0002,
      "per_gb_egress_usd": 0.05
    },
    "accepts": [{ "chain": "base", "asset": "USDC", "recipient": "0x15184b..." }],
    "expires_at": 1777220000,
    "tier": 2,
    "product": "cdn_prepaid_bundle"
  }
}

Submit proof to /v1/x402/proof/submit:

{ "nonce": "...", "payer": "0x...", "chain": "base", "tx_hash": "0x..." }

…then call /v1/cdn/get with X-Hive-Access: hive_<token>. Each request decrements the prepaid balance by per_request + (bytes / 1GB) * per_gb_egress. The residual balance rides the X-Hive-Balance-USD response header.

Cache semantics

  • Cache-Control: max-age=<s> on PUT sets the TTL. Default 300 seconds. Capped at 7 days.
  • ETag is a strong validator (SHA-256 prefix in quotes).
  • If-None-Match: <etag> on GET returns 304 Not Modified with no body and no egress charge.
  • Expired entries are dropped lazily on read and on a 60-second sweep.
  • A PUT over an existing key replaces the body, ETag, content type, TTL, and resets the hit counter.

Environment

Var Default Purpose
PORT 3000 HTTP port.
ENABLE true Master switch. false runs /health only.
WALLET_ADDRESS 0x15184bf50b3d3f52b60434f8942b7d52f2eb436e Recipient for x402 fees.
CDN_PRICE_PER_REQUEST_USD 0.0002 Per-request fee.
CDN_PRICE_PER_GB_EGRESS_USD 0.05 Per-GB egress fee.
CDN_PREPAY_BUNDLE_USD 1.0 Default prepay bundle size.
CDN_DEFAULT_TTL_S 300 Default TTL for objects without explicit max-age.
CDN_LRU_MAX_ENTRIES 1024 Hot-tier entry cap.
CDN_LRU_MAX_BYTES 67108864 Hot-tier byte cap (64 MiB).
CDN_MAX_OBJECT_BYTES 4194304 Per-object size limit (4 MiB).
CDN_DB_PATH /tmp/cdn.db SQLite warm-tier path.

What this server will not do

  • Make outbound requests of any kind. The cache only holds what was put into it.
  • Resolve URLs, fetch origins, or revalidate against a backing store.
  • Enforce per-key access control. Treat keys as opaque namespaces and gate them upstream if you need privacy.
  • Persist beyond the SQLite file. On a fresh container with a new /tmp, the cache is empty.

License

MIT.

<!-- HIVE-GAMIFICATION-META-START -->

Hive Gamification

This MCP server is part of the Hive Civilization gamification surface (10-mechanic capability taxonomy).

  • Capability taxonomy: https://hive-gamification.onrender.com/.well-known/hive-gamification.json
  • Centrifuge dashboard: https://hive-gamification.onrender.com/.well-known/hive-centrifuge.json
  • Consolidated OpenAPI: https://hive-gamification.onrender.com/.well-known/openapi.json

Surface tags: gamification.spec.v1 · gamification.surface.public · gamification.signal.read-only · gamification.settlement.real-rails

Real rails on Base L2 (USDC 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913). Read-only signal layer. Brand gold #C08D23. <!-- HIVE-GAMIFICATION-META-END -->

Hive Civilization Directory

Part of the Hive Civilization — agent-native financial infrastructure.

  • Endpoint Directory: https://thehiveryiq.com
  • Live Leaderboard: https://hive-a2amev.onrender.com/leaderboard
  • Revenue Dashboard: https://hivemine-dashboard.onrender.com
  • Other MCP Servers: https://github.com/srotzin?tab=repositories&q=hive-mcp

Brand: #C08D23 <!-- /hive-footer -->

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured