Hive Agent Storage

io.github.srotzinPublished by GitHub account srotzinv1.0.0Updated Oct 24, 2026

Encrypted A2A object storage for autonomous agent state and artifacts

The publisher's own server: Tigris MCP Server

VerifiedStreamable HTTPWeb executableFiles & StorageFinance

Overview

AI-generated overview

Remote MCP server for agent-scoped object storage with DID-isolated namespaces and per-KB USDC metering, though its storage backend is not yet live.

What you can do
  • Upload, read, list and tombstone objects in an agent-isolated namespace, with hot, warm or cold retention classes.
  • Check quota: bytes used and allocated, object count and lifetime spend.
  • Use the shared earn rails to register, view your own record and see a leaderboard.
  • Discover the service through A2A AgentCard and MCP discovery endpoints.
When to use it
Worth considering only if you want an agent to store state or artifacts in a DID-scoped namespace and are prepared to pay per kilobyte in USDC. As shipped, the storage tools return a backend-pending error, so it is not usable for real storage yet.
Requirements
A remote Streamable-HTTP endpoint; no local package or runtime. An API key bound to the agent's DID at the provider's gate is required for reads and writes inside the namespace. Paid calls settle in USDC on the Base network, so a funded wallet is needed.
Before you install
Writes and cross-DID reads cost real USDC on Base, not testnet funds; the README states the storage backend is not live and storage tools return a 503-style error. An API key bound to a DID is required. Cold-tier deletes unlink from the namespace but the object remains on-chain. The README also lists a fixed settlement recipient address.

Installation

In SourceWeft

  1. Open Hive Agent Storage in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.srotzin%2Fhive-mcp-agent-storage/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "hive-agent-storage": {
      "type": "http",
      "url": "https://hive-mcp-agent-storage.onrender.com/mcp"
    }
  }
}

README

[Hive Agent Storage · Agent-Native Object Storage MCP]

hive-mcp-agent-storage

Agent-native object storage MCP server. Per-agent DID isolation. x402 pay-per-byte. Real Base USDC.

[Smithery] [Glama] [Backend] [Release] [License]

https://hive-mcp-agent-storage.onrender.com/mcp


Hive Agent Storage

Agent-native object storage with per-agent DID isolation and x402 pay-per-byte metering.

hive-mcp-agent-storage is a Model Context Protocol (MCP) server that gives autonomous agents a first-class object-storage primitive. Each agent gets its own DID-isolated namespace; reads inside the namespace are free, writes are metered per KB in real Base USDC, and cross-DID reads are metered per KB. Storage is routed under the hood to Storj (hot), Filecoin (warm), or Arweave (cold) depending on retention class.

The differentiator is not a new storage backend. It is the agent-native shape: DID-scoped namespaces, x402 settlement, content-addressed receipts, and an MCP/A2A-compatible discovery surface.

  • Protocol: MCP 2024-11-05 over Streamable-HTTP / JSON-RPC 2.0
  • A2A: AgentCard at /.well-known/agent.json
  • x402: every paid call settles in real Base USDC to 0x15184bf50b3d3f52b60434f8942b7d52f2eb436e
  • Rails: USDC on Base L2 — no mocks, no testnet, no dev-trust
  • Author: Steve Rotzin · Hive Civilization · brand gold #C08D23 (Pantone 1245 C)

Backend status — pending

The MCP shim, registry manifests, AgentCard, and JSON-LD ship today as a discovery anchor. The upstream backend at hivemorph.onrender.com/v1/storage/* is not yet live. Until it ships, every storage tool returns a 503-shape body:

json
{  "error": "backend_pending",  "retry_after": 86400,  "detail": "Hive Agent Storage backend endpoints (/v1/storage/*) are not yet live...",  "brand": "#C08D23"}

The shared hive_earn_* tools route to live /v1/earn/* endpoints and behave normally.

Tools

ToolPurposePricing
agent_storage_putUpload bytes into an agent-isolated namespace. Choose hot/warm/cold retention class.$0.0001/KB upload, real Base USDC
agent_storage_getRead an object. Free for own DID.$0.00005/KB cross-DID, real Base USDC
agent_storage_listList object keys with prefix + pagination.Free
agent_storage_deleteTombstone an object (owner-only).Free
agent_storage_quotaBytes used / allocated / object count / lifetime spend.Free

Plus the standard Hive earn rails: hive_earn_register, hive_earn_me, hive_earn_leaderboard.

Endpoints

PathPurpose
POST /mcpJSON-RPC 2.0 / MCP 2024-11-05
GET /HTML landing with comprehensive meta tags + JSON-LD
GET /healthHealth + telemetry
GET /.well-known/mcp.jsonMCP discovery descriptor
GET /.well-known/agent.jsonA2A AgentCard
GET /.well-known/oac.jsonOpen Agent Card JSON-LD
GET /.well-known/security.txtRFC 9116 security contact
GET /robots.txtAllow-all crawl policy
GET /sitemap.xmlCrawler sitemap
GET /og.svg1200×630 Hive-gold OG image
GET /seo.jsonJSON-LD structured data (SoftwareApplication)

Connect

Claude Desktop / Cursor / Manus and other MCP clients can mount the server via Streamable-HTTP:

json
{  "mcpServers": {    "hive-agent-storage": {      "url": "https://hive-mcp-agent-storage.onrender.com/mcp"    }  }}

Smithery one-click install: https://smithery.ai/server/hivecivilization (server: hive-mcp-agent-storage).

Architecture

       MCP / A2A clients            │            ▼  hive-mcp-agent-storage  (this repo)            │  per-agent DID isolation            │  x402 metering shim            ▼   /v1/storage/*  (hivemorph.onrender.com — pending)            │            ├─ Storj      (hot, frequent access)            ├─ Filecoin   (warm, periodic access)            └─ Arweave    (cold, permanent)

Per-agent DID isolation means:

  • The agent's DID is the namespace key. Reads/writes inside the namespace require an API key bound to that DID at HiveGate.
  • Cross-DID reads are explicitly metered and audited; there is no implicit trust.
  • Tombstone receipts are chain-attested; cold-tier (Arweave) deletes unlink from the namespace but the object remains on-chain by design.

License

MIT. © Steve Rotzin / Hive Civilization. Brand gold #C08D23 (Pantone 1245 C). Never #f5c518.

Agent-native (v1.0.0)

This shim ships the Hive Civilization agent-native bundle so any A2A or MCP-aware agent can discover, pay, and earn:

  • A2A AgentCard — GET /.well-known/agent.json (also at /agent.json).
  • Open Agent Card (OAC) JSON-LD — embedded inline at / and /agent.html, with @type SoftwareApplication + @type AgentCard under @context https://schema.org + https://a2a-protocol.org/v1.
  • Earn rails — hive_earn_register, hive_earn_me, hive_earn_leaderboard against https://hivemorph.onrender.com/v1/earn/*. Resilient to upstream cold-start.
  • x402 propagation — paid responses pass through the upstream 402 body untouched so the consuming agent can auto-pay.
  • Pricing annotations — every paid tool descriptor carries a non-standard pricing block (amount / currency / chain / recipient) ahead of MCP-next.
  • Brand: Hive Civilization gold #C08D23. Settlement: real Base USDC, recipient 0x15184bf50b3d3f52b60434f8942b7d52f2eb436e. No mock, no testnet.

Hive Civilization Directory

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

Brand: #C08D23

Source: README.md at commit d3f4112

Tools

0
SourceWeft has not read this server's tools yet.

Version history

1
  1. v1.0.0LatestSep 16, 2026

Score

85% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listedRequired15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

Other Storage integrations