M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
H

Hive Intelligence MCP

Community
Managed crypto market infrastructure for AI agents
How to install ↓ Compare GitHub source repository ↗
Category
Other #96 of 230
Stars
★ 20 Popular
Transport
Streamable HTTP · stdio (local process)
Runtime
Node.js · Hosted (no local runtime)
Credentials
API key / credential required
License
MIT
Last commit
Tools
6
57FMRS · C

Hive Intelligence MCP aggregates nine crypto data sources behind one API key with 369 tools and per-call provenance, well suited for agents needing broad crypto context; the server itself is closed-source and managed, priced by credits, and its security outputs should be interpreted cautiously.

Strongest · Documentation 16/20 Weakest · Reliability 6/20

Reliability
6/20
Security and permissions
12/20
Maintenance
10/20
Documentation
16/20
Setup experience
13/20
Why each score
Reliability 6/20
The repository does not contain the actual running MCP server code — the README explicitly states that mcp.hiveintelligence.xyz is a 'managed, proprietary service,' with only the client SDK and skills being open source. This means the claimed capabilities (369 tools, 10 categories, etc.) cannot be verified against this repo's contents, and whether declared tools match real behavior is entirely unverifiable statically. No CI workflows or committed tests covering the stdio/streamable-http paths declared in server.json are present. Per the static calibration rule, without verifiable execution evidence reliability cannot exceed 12; given the core server code is absent and behavior is wholly unverifiable, 6 is awarded.
Security and permissions 12/20
Credential handling shows positive signals: all API key environment variables (COINGECKO_PRO_API_KEY, ALCHEMY_API_KEY, etc.) are marked isSecret:true and isRequired:false in the manifest, the remote endpoint's Authorization/X-API-Key headers are likewise optional secret fields, and README examples use placeholders rather than real tokens, so no red line is triggered. The declared tools (e.g. detect_rugpull, get_token_security, transaction simulation) are read-only/pre-signing checks, with no evidence of irreversible actions (transfers, deploys, signing) executed by default. However, because the actual hosted server is closed-source and proprietary, claims like 'never silently mixes provider data' or how keys are actually processed cannot be verified in code — they rest solely on the README (untrusted evidence). The stdio npm package (hive-intelligence) source is also not included in the review material, so its outbound network behavior likewise cannot be confirmed. Given visible primary risks but unverifiable permission/data-boundary claims, 12 is awarded.
Maintenance 10/20
The repo carries an MIT license, is not archived, and the manifest shows a tracked version (1.5.3), indicating some versioning discipline. But the provided evidence lacks commit history, release cadence, issue-response times, or a dedicated security-response channel (e.g. no SECURITY.md); the project is small (18 stars, 3 open issues). Support channels are general (email/Telegram/Twitter) rather than security-specific. This reads as 'active but with governance/versioning evidence gaps,' warranting 10.
Documentation 16/20
The README is thorough, covering connection configs for multiple clients (Claude Code, Cursor, VS Code, Claude Desktop, Windsurf, Gemini CLI), authentication guidance, an environment-variable table, a pricing table, CLI command reference, a tool-category overview, a security & trust section, and an FAQ — and it transparently discloses that the core server is proprietary rather than overstating openness, which is a credibility plus. Deductions apply because detailed tool parameters/schemas are not embedded in the repo docs (requiring live endpoint discovery), and rate-limit/cost specifics rely mostly on external links rather than in-repo evidence. 16 is awarded.
Setup experience 13/20
The installation path is clear: hosted mode needs only a Bearer/X-API-Key header to connect via several mainstream clients, with per-client config snippets provided; local stdio mode runs via npx against a published npm package. Steps are few and well-specified. Per the static calibration rule, however, without CI workflows and committed tests covering key paths, setup cannot exceed 15; additionally, since neither the hosted server nor the stdio npm package's source is in this repo, 'a clean start plus successful init/tool-list handshake' cannot be verified, and full verification would still require external account/API-key registration. 13 is awarded.

Static review · not runListed 2026-09-07

Read the FMRS scoring method →

Fit and risk

What it can accessUses the network

Best for

  • AI agent developers needing broad crypto context — prices, DeFi, wallets, security — from multiple providers in one place
  • Teams doing crypto due diligence with Claude, Cursor, VS Code, or other remote-HTTP-MCP clients
  • Integrators who want one API key instead of separate upstream provider credentials

Not for

  • Users who need to self-host or modify the MCP server source (the server is closed-source and managed)
  • Users seeking unlimited free usage (credit-based pricing applies)
  • General-purpose tasks unrelated to crypto

Required permissions

  • Hive API key (Authorization Bearer or X-API-Key header, optional secret)
  • Optional upstream provider keys (COINGECKO_PRO_API_KEY, ALCHEMY_API_KEY, HELIUS_API_KEY, MORALIS_API_KEY, TENDERLY_API_KEY, GOPLUS_APP_KEY/SECRET, CODEX_API_KEY, etc.)

Risks and side effects

  • Credit-based billing: each tool call consumes credits and rate limits apply
  • Security scores are not investment advice; honeypot/rugpull results may be incomplete or lag
  • On-chain text (token names, memos) should be treated as untrusted input with prompt-injection risk
  • Data may come from cached or degraded sources; check provenance and freshness labels in responses
  • Leaked API keys can be abused; keep keys server-side and rotate them

Setup

Before you start

Runtime：Node.js · Hosted (no local runtime)

HIVE_API_KEY requiredsecret Required Hive API key, obtained at hiveintelligence.xyz/dashboard/keys or via hive auth login.
COINGECKO_PRO_API_KEY optionalsecret Optional CoinGecko Pro API key for upstream data when self-hosting.
ALCHEMY_API_KEY optionalsecret Optional Alchemy API key (EVM wallet, token, NFT, simulation, gas, RPC data).
CODEX_API_KEY optionalsecret Optional Codex GraphQL API key.
GOPLUS_APP_KEY optionalsecret Optional GoPlus application key.
GOPLUS_APP_SECRET optionalsecret Optional GoPlus application secret paired with the app key.
HELIUS_API_KEY optionalsecret Optional Helius API key (Solana RPC and DAS data).
MORALIS_API_KEY optionalsecret Optional Moralis API key (EVM wallet, token, NFT analytics).
TENDERLY_API_KEY optionalsecret Optional Tenderly API key (EVM simulation, tracing, gas estimation).
TENDERLY_NODE_ACCESS_KEY optionalsecret Optional Tenderly Node access key (EVM RPC and historical state queries).
TENDERLY_GATEWAY_ACCESS_KEY optionalsecret Optional legacy alias for the Tenderly Node access key.
Other optional settings (4)
HIVE_API_URL optional Optional custom base URL, defaults to https://mcp.hiveintelligence.xyz.
API_EXECUTE_ENDPOINT optional Advanced override for the execute endpoint.
TENDERLY_ACCOUNT_SLUG optional Optional Tenderly account slug for project-scoped API calls.
TENDERLY_PROJECT_SLUG optional Optional Tenderly project slug for project-scoped API calls.
  1. Get an API key at hiveintelligence.xyz/dashboard/keys (free Demo plan, no card required).
  2. Hosted endpoint: run claude mcp add --transport http hive https://mcp.hiveintelligence.xyz/mcp --header "Authorization: Bearer hive_live_..." in Claude Code, or add the url + headers block to mcp. for Cursor, VS Code, Windsurf, or Gemini CLI, replacing the placeholder key. For Claude Desktop, use Settings → Connectors → Add custom connector with the URL and Authorization header.
  3. Local stdio: npx -y -p hive-intelligence@latest hive with the HIVE_API_KEY environment variable.
claude_desktop_config.json 
{"mcpServers":{"hive":{"url":"https://mcp.hiveintelligence.xyz/mcp","headers":{"Authorization":"Bearer hive_live_..."}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json
{
  "servers": {
    "hive": {
      "url": "https://mcp.hiveintelligence.xyz/mcp",
      "headers": {
        "Authorization": "Bearer hive_live_..."
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal
claude mcp add --transport http hive https://mcp.hiveintelligence.xyz/mcp --header 'Authorization: Bearer hive_live_...'

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After connecting, tools like search_tools and get_price should appear in your client's tool list; running search_tools or asking "What's the price of bitcoin right now?" should return a live, source-attributed result, confirming the connection works.

Troubleshooting

  1. Verify the API key is valid, credits remain, and the plan's rate limit is not exceeded
  2. If the remote connection fails, confirm the client supports Streamable HTTP and API-key headers; Claude Desktop requires Custom Connectors, not claude_desktop_config.
  3. VS Code configs must explicitly set type: "http"
  4. For local stdio, run hive doctor or hive status; confirm HIVE_API_KEY is set
  5. If data looks wrong, check the provider attribution and timestamps on the response

Things to try

Once connected, you can ask your AI assistant things like:

  • What's the price of BTC, ETH, and SOL right now in USD?
  • List the top 20 yield pools above 10% APY on Ethereum.
  • Is this token a honeypot? 0x...
  • Show me the portfolio of vitalik.eth across all chains.

Tools 6

search_tools read-only
Discover the right task toolset for the agent
get_api_endpoint_schema read-only
Inspect endpoint parameter schemas
invoke_api_endpoint read-only
Run a bounded API call
get_token_security read-only
Fetch token security and honeypot risk flags
detect_rugpull read-only
Run a rugpull check on a token
get_price read-only
Query crypto prices

Use cases

Query live BTC/ETH/SOL prices and OHLCV
List Ethereum yield pools above 10% APY
Inspect vitalik.eth's portfolio across all chains
Run honeypot detection, rugpull checks, and Tenderly transaction simulation before signing
Track trending Polymarket events and perp funding rates

Supported clients

Claude Desktop
Claude Code
Cursor
VS Code
Windsurf
Gemini CLI

Listed from the project's documentation, not tested by this site.

Overview

Hive Intelligence MCP is a hosted crypto data MCP server that normalizes nine upstream providers — Alchemy, CoinGecko, DeFiLlama, Moralis, Codex, GoPlus, Helius, Tenderly, and CCXT — into 369 tools across 10 categories covering market data, DeFi, wallets, token security, DEX flows, NFTs, prediction markets, and on-chain data, all behind a single API key. Every call returns provider attribution, freshness, and runtime status. It offers a hosted streamable-http endpoint and a local stdio mode via the npm package hive-intelligence. The SDK and agent skills are MIT open source; the MCP server itself is a managed, proprietary service.

Similar servers

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 659 · Tools 11 Compare with this →
Compare with other MCP servers

Source revision 30cfea8b7e36 Data synced 2026-10-19 Read the FMRS scoring method