# Agents — Hive Intelligence

> Machine-readable manifest for AI agents installing or evaluating Hive Intelligence.
> Companion to /llms.txt (broader site context) and /agent-onboarding/SKILL.md (full install path tree).
>
> Last updated: 2026-06-06.

## What Hive does for an agent

Hive Intelligence is a managed MCP server, REST API, and CLI that gives AI agents structured, runtime access to live cryptocurrency market data, DeFi activity, wallet positions, token security signals, DEX flows, prediction markets, NFT analytics, and on-chain network data — all through one normalized tool surface, one API key, and one credit currency.

On root `/mcp`, agents call `tools/list` for the compact 13-tool discovery layer, then use `search_tools`, `get_api_endpoint_schema`, and `invoke_api_endpoint` to call exact provider tools. New tools appear in `hive://tools` and authenticated REST `https://mcp.hiveintelligence.xyz/api/v1/tools` as Hive expands coverage, so agents can inspect current schemas at runtime instead of hand-maintaining provider wrappers.

## Canonical install paths

Pick one of four paths. The recommended path depends on what the agent already has access to:

1. **Live query (no install)** — Use the public Playground for one-off questions: https://www.hiveintelligence.xyz/playground
2. **App integration** — Add the MCP server to a supported AI client or API path (Claude Desktop, Claude Code, Cursor, Windsurf, VS Code, OpenAI Responses API, Codex CLI, Gemini CLI). Per-client/API guides at https://www.hiveintelligence.xyz/install
3. **Get an API key** — For REST or programmatic use. Sign in at https://www.hiveintelligence.xyz/login?next=/dashboard/keys and create a dashboard key
4. **CLI setup** — `npx -y -p hive-intelligence@latest hive init --all --browser` runs the Hive CLI setup wizard for local credentials and supported local agent resources. Use the per-client/API guides for integration-specific MCP details.

## MCP endpoints

- Root MCP: `https://mcp.hiveintelligence.xyz/mcp` (POST, JSON-RPC 2.0, stateless Streamable HTTP)
- Category endpoints (one per category, see table below)
- REST execute: `POST https://mcp.hiveintelligence.xyz/api/v1/execute` with body `{ "tool": "tool_name", "args": { ... } }`
- Authenticated tool discovery: `GET https://mcp.hiveintelligence.xyz/api/v1/tools`
- B2B readiness: `GET https://mcp.hiveintelligence.xyz/api/v1/b2b/readiness`
- OpenAPI spec: `GET https://mcp.hiveintelligence.xyz/api/openapi.json`
- Task canary health: `GET /api/health/toolsets`, `GET /api/health/toolsets/quick`, `GET /api/health/toolsets/history`
- Workflow regression health: `GET /api/health/workflows`, `GET /api/health/workflows/quick`, `GET /api/health/workflows/history`
- MCP discovery descriptor: `GET https://www.hiveintelligence.xyz/.well-known/mcp.json`
- MCP server card: `GET https://www.hiveintelligence.xyz/.well-known/mcp/server-card.json`

Before approving a client or marketplace connection, verify that the server URL
is Hive's official endpoint: `https://mcp.hiveintelligence.xyz/mcp`. Do not approve a
look-alike MCP URL that points somewhere else.

## Authentication

Send one of these headers with every request:

- `Authorization: Bearer <YOUR_HIVE_API_KEY>` (recommended)
- `x-api-key: <YOUR_HIVE_API_KEY>` (legacy alias, still accepted)

Create or select a key at https://www.hiveintelligence.xyz/login?next=/dashboard/keys. Plan limits control monthly credits and rate limits; use https://www.hiveintelligence.xyz/pricing when the user asks about budgets or throughput.

## Tool catalog (10 categories)

| Category | MCP Endpoint | Marketing URL | Domain |
| --- | --- | --- | --- |
| Market Data & Price | `/hive_market_data/mcp` | `/tools/market-data` | Spot, OHLC, market caps, volume, derivatives, exchanges. |
| On-Chain DEX & Pool Analytics | `/hive_onchain_dex/mcp` | `/tools/dex-analytics` | Pool, pair, liquidity, and trade flow data across major EVM and Solana markets. |
| Portfolio & Wallet | `/hive_portfolio_wallet/mcp` | `/tools/portfolio` | Wallet balances, history, PnL, approvals across EVM and Solana. |
| Token & Contract Data | `/hive_token_contract/mcp` | `/tools/token-contract` | Token metadata, holders, top traders, lifecycle events. |
| DeFi Protocol Analytics | `/hive_defi_protocol/mcp` | `/tools/defi` | TVL, yields, fees, chain breakdowns, stablecoins. |
| NFT Analytics | `/hive_nft_analytics/mcp` | `/tools/nft` | Collections, floors, sales, holders, traits, compressed NFTs. |
| Security & Risk Analysis | `/hive_security_risk/mcp` | `/tools/security` | Token, NFT, dApp, wallet scans; honeypot, rug-pull, simulation, LP locks. |
| Network & Infrastructure | `/hive_network_infrastructure/mcp` | `/tools/network` | Network lists, gas, block numbers, exchange status, and Solana infrastructure. |
| Search & Discovery | `/hive_search_discovery/mcp` | `/tools/search` | Cross-provider search across coins, tokens, exchanges, categories, events. |
| Prediction Markets | `/hive_prediction_markets/mcp` | `/tools/prediction-markets` | Polymarket events, traders, market prices, settlement history. |

The public overview of the wrapped provider catalog is at https://www.hiveintelligence.xyz/tools/live-catalog, and the runtime canonical list is served through authenticated REST at https://mcp.hiveintelligence.xyz/api/v1/tools.

## Canonical task toolsets

Use task toolsets as the agent-facing selection layer before choosing exact tools. They do not remove, hide, or replace provider tools; all registered provider tools remain callable through `invoke_api_endpoint`, the REST execute endpoint, and category MCP endpoints.

Runtime source: read `hive://toolsets` from the root MCP endpoint. Discovery pattern: `search_tools` → `get_api_endpoint_schema` → `invoke_api_endpoint`.

| Toolset ID | Label | Kind | Purpose | Outcome | Required identifiers | Recommended tools |
| --- | --- | --- | --- | --- | --- | --- |
| `market_research` | Market Research | product | Answer price, market cap, OHLC, exchange, order book, derivatives, and trading-context questions. | A current market view with exact asset/exchange fixtures, price context, liquidity, volume, and freshness/status notes. | token symbol or contract, chain or exchange id | `get_price`, `get_coins_market_data`, `get_coin_market_chart_range`, `get_coin_ohlc`, `get_coin_tickers`, `get_ticker` |
| `token_diligence` | Token Diligence | product | Investigate metadata, market context, liquidity, holders, DEX pairs, risk, and enrichment signals. | A replayable token report separating market data, on-chain activity, liquidity, holder concentration, and risk findings. | chain, token contract or exact token id | `search_tokens`, `get_token_info`, `get_token_price`, `get_token_chart_data`, `get_token_holders`, `get_token_pairs` |
| `wallet_investigation` | Wallet Investigation | product | Inspect balances, transfers, PnL, NFTs, DeFi positions, chain activity, and decoded history. | A wallet exposure and activity report with holdings, notable transfers, DeFi/NFT context, and missing-data status. | wallet address, chain | `alchemy_get_tokens_by_wallet`, `alchemy_get_token_balances_by_wallet`, `alchemy_get_nfts_by_wallet`, `alchemy_get_asset_transfers`, `alchemy_get_native_balance`, `moralis_get_wallet_history` |
| `security_risk` | Security Risk | product | Check token, address, approval, contract, dApp, phishing, and simulation risk signals. | A risk report with severity, exact evidence, runtime status, and remediation-oriented next checks. | chain, token contract, wallet address, or transaction payload | `get_token_security`, `get_wallet_approvals`, `check_malicious_address`, `check_phishing_site`, `check_dapp_security`, `check_approval_security` |
| `onchain_dex_pool_analysis` | On-chain DEX and Pool Analysis | product | Analyze pools, pairs, liquidity, trades, OHLCV, trending DEX activity, and token-level DEX flow. | A DEX/pool report covering liquidity, trading activity, pair quality, and comparable on-chain context. | chain, pair address or token contract | `get_trending_pools`, `get_pool_info`, `get_pool_ohlcv`, `get_pool_trades`, `get_token_trades`, `get_token_pools` |
| `defi_protocol_analysis` | DeFi Protocol Analysis | product | Analyze protocols, TVL, fees, revenue, stablecoins, bridges, yields, and aggregate DeFi metrics. | A protocol or DeFi-sector report with TVL, fee/revenue, yield, bridge/stablecoin, and chain-distribution context. | protocol name or chain | `get_protocol_tvl`, `get_protocol_fees`, `get_protocol_fee_summary`, `get_chain_fees`, `get_global_defi`, `get_yield_pools` |
| `nft_research` | NFT Research | product | Research NFT ownership, metadata, collection context, floors, sales, rarity, spam checks, and Solana assets. | An NFT or collection report with owner/metadata, market context, authenticity/spam signals, and provenance fixtures. | chain, collection contract or asset id | `alchemy_get_nfts_for_owner`, `alchemy_get_nfts_for_contract`, `alchemy_get_nft_metadata`, `alchemy_get_floor_price`, `alchemy_get_nft_sales`, `alchemy_get_owners_for_contract` |
| `network_infrastructure` | Network Infrastructure | product | Read chain state, gas, blocks, receipts, logs, transaction status, supported networks, and Solana priority fees. | A chain/RPC diagnostic answer with exact network, block/transaction fixtures, and execution-adjacent status. | chain | `alchemy_get_block_number`, `alchemy_get_chain_id`, `alchemy_get_gas_price`, `alchemy_get_transaction_by_hash`, `alchemy_get_transaction_receipt`, `alchemy_get_logs` |
| `solana_analysis` | Solana Analysis | product | Analyze Solana-native wallets, SPL token accounts, DAS assets, parsed transactions, priority fees, and launch data. | A Solana-specific wallet/token/asset report using Solana-native identifiers and runtime status. | Solana wallet, mint, asset id, or transaction signature | `helius_get_balance`, `helius_get_token_balances`, `helius_get_assets_by_owner`, `helius_get_token_accounts`, `helius_get_parsed_transactions`, `helius_get_parsed_address_transactions` |
| `prediction_markets` | Prediction Markets | product | Research prediction markets, events, market stats, prices, outcomes, traders, holders, and trades. | A prediction-market report with exact market/event identifiers, price/stats, trader context, and outcome notes. | market id, event id, or search query | `codex_filter_prediction_markets`, `codex_event_scoped_filter_prediction_markets`, `codex_prediction_markets`, `codex_prediction_market_price`, `codex_prediction_outcome_order_books`, `codex_prediction_market_stats` |
| `stateful_monitoring` | Stateful Monitoring | product | Create and manage durable crypto intelligence monitors that remember user intent across agent sessions. | A saved monitor configuration for wallets, tokens, protocols, markets, prediction markets, watchlist digests, token discovery risk, or risk watch that workers can schedule, compare, alert on, and report from. | monitor kind, target object | `hive_create_monitor`, `hive_list_monitors`, `hive_get_monitor`, `hive_update_monitor`, `hive_archive_monitor`, `hive_get_monitor_runs` |
| `search_discovery` | Search and Discovery | meta | Route ambiguous requests to exact task toolsets, categories, tools, schemas, and entity identifiers. | A narrowed execution plan with the right toolset, exact tools, required arguments, and fixtures to request. | natural-language query | `search_tools`, `get_api_endpoint_schema`, `invoke_api_endpoint`, `search_all`, `search_tokens`, `search_pools` |

## Task canaries and runtime status

`hive://task-canaries` lists configured task probes. The HTTP health surface can execute or read them:
`GET /api/health/toolsets`, `GET /api/health/toolsets/quick`,
`GET /api/health/toolsets/history`, and `GET /api/health/toolsets/:toolsetId`.

| Toolset ID | Probe tool | Purpose | Fixture args |
| --- | --- | --- | --- |
| `market_research` | `get_price` | CoinGecko BTC/USD price sanity check. | `{"ids":"bitcoin","vs_currencies":"usd"}` |
| `token_diligence` | `alchemy_get_token_metadata` | Alchemy USDC metadata check for token diligence. | None |
| `wallet_investigation` | `alchemy_get_native_balance` | Alchemy public wallet native balance check. | None |
| `security_risk` | `get_token_security` | GoPlus USDT token security check. | None |
| `onchain_dex_pool_analysis` | `get_pool_info` | GeckoTerminal USDC/WETH pool info check. | None |
| `defi_protocol_analysis` | `get_protocol_tvl` | DeFiLlama Aave TVL check. | None |
| `nft_research` | `alchemy_get_contract_metadata` | Alchemy BAYC collection metadata check. | None |
| `network_infrastructure` | `alchemy_get_block_number` | Alchemy Ethereum latest block check. | None |
| `solana_analysis` | `helius_get_slot` | Helius Solana slot check. | None |
| `prediction_markets` | `codex_filter_prediction_markets` | Codex open prediction market discovery check. | None |
| `stateful_monitoring` | `hive_list_monitors` | Hive durable monitor store availability check. | `{"__hive_user_id":"00000000-0000-4000-8000-000000000000"}` |
| `search_discovery` | `search_all` | Search/discovery query check. | `{"query":"bitcoin"}` |

| Runtime status | Meaning |
| --- | --- |
| `ok` | Provider call is expected to work with current configuration. |
| `missing_key` | Provider credentials are not configured for this runtime. |
| `plan_required` | The tool exists but the upstream account plan, quota, or credits prevent execution. |
| `rate_limited` | The upstream or Hive rate limit was hit. Retry after backoff or Retry-After. |
| `degraded` | The tool exists, but the provider is temporarily unavailable or unstable. |
| `failing` | The tool is returning an unexpected implementation or upstream error. |

## Hive skills pack

Skills are procedural guidance for agents. They do not execute data calls by
themselves; Hive MCP remains the execution layer. Runtime source: read
`hive://skills`. Public JSON mirror: https://www.hiveintelligence.xyz/skills/index.json.
Installable skills mirror: https://github.com/hive-intel/hive-skills
(`npx skills add hive-intel/hive-skills`). List available skills with
`npx skills add hive-intel/hive-skills --list`.

| Skill | Name | Toolsets | Purpose |
| --- | --- | --- | --- |
| [Install Hive MCP](https://www.hiveintelligence.xyz/skills/hive-mcp.md) | `hive-mcp` | `search_discovery` | Connect Hive's hosted remote MCP endpoint to Claude Desktop, Claude Code, Cursor, Windsurf, VS Code, OpenAI Responses API, Codex CLI, and Gemini CLI. |
| [Get a Hive API Key](https://www.hiveintelligence.xyz/skills/hive-build-onboarding.md) | `hive-build-onboarding` | `search_discovery` | Guide agents through browser auth, dashboard key creation, and headless environment setup. |
| [Build With Hive](https://www.hiveintelligence.xyz/skills/hive-build.md) | `hive-build` | `search_discovery` | Use Hive from application code through REST, MCP clients, or the published TypeScript adapter. |
| [Use Hive CLI](https://www.hiveintelligence.xyz/skills/hive-cli.md) | `hive-cli` | `search_discovery` | Run Hive from shell scripts, cron jobs, and terminal research workflows. |
| [Route A Crypto Query](https://www.hiveintelligence.xyz/skills/hive-query.md) | `hive-query` | `market_research`, `token_diligence`, `wallet_investigation`, `security_risk`, `onchain_dex_pool_analysis`, `defi_protocol_analysis`, `nft_research`, `network_infrastructure`, `solana_analysis`, `prediction_markets`, `stateful_monitoring`, `search_discovery` | Start from the user's intent, choose the canonical task toolset, inspect schemas, and call exact Hive tools. |
| [Stateful Monitoring](https://www.hiveintelligence.xyz/skills/hive-stateful-monitoring.md) | `hive-stateful-monitoring` | `stateful_monitoring` | Create, list, update, and archive durable crypto intelligence monitors that agents can schedule, remember, and report from. |
| [Tool Discovery](https://www.hiveintelligence.xyz/skills/hive-tool-discovery.md) | `hive-tool-discovery` | `search_discovery` | Use search_tools, hive://toolsets, schema lookup, and invoke_api_endpoint correctly. |
| [Market Research](https://www.hiveintelligence.xyz/skills/hive-market-research.md) | `hive-market-research` | `market_research` | Research prices, liquidity, exchanges, OHLC, order books, and derivatives context. |
| [Token Diligence](https://www.hiveintelligence.xyz/skills/hive-token-diligence.md) | `hive-token-diligence` | `token_diligence` | Investigate token metadata, market context, liquidity, holders, enrichment, and risk. |
| [Wallet Investigation](https://www.hiveintelligence.xyz/skills/hive-wallet-investigation.md) | `hive-wallet-investigation` | `wallet_investigation` | Inspect wallet balances, transfers, NFTs, PnL, DeFi positions, and notable activity. |
| [Security Risk](https://www.hiveintelligence.xyz/skills/hive-security-risk.md) | `hive-security-risk` | `security_risk` | Check token, approval, address, phishing, contract, and simulation risk signals. |
| [DEX And Pool Analysis](https://www.hiveintelligence.xyz/skills/hive-dex-pool-analysis.md) | `hive-dex-pool-analysis` | `onchain_dex_pool_analysis` | Analyze on-chain pools, pairs, liquidity, trades, OHLCV, and DEX flow. |
| [DeFi Research](https://www.hiveintelligence.xyz/skills/hive-defi-research.md) | `hive-defi-research` | `defi_protocol_analysis` | Analyze protocols, TVL, fees, revenue, stablecoins, bridges, and yield pools. |
| [NFT Research](https://www.hiveintelligence.xyz/skills/hive-nft-research.md) | `hive-nft-research` | `nft_research` | Research NFT collections, ownership, metadata, floor price, sales, rarity, and spam checks. |
| [Solana Analysis](https://www.hiveintelligence.xyz/skills/hive-solana-analysis.md) | `hive-solana-analysis` | `solana_analysis` | Analyze Solana wallets, SPL token accounts, DAS assets, parsed transactions, and priority fees. |
| [Network Infrastructure](https://www.hiveintelligence.xyz/skills/hive-network-infrastructure.md) | `hive-network-infrastructure` | `network_infrastructure` | Read chain state, gas, blocks, receipts, logs, transaction status, and RPC diagnostics. |
| [Prediction Markets](https://www.hiveintelligence.xyz/skills/hive-prediction-markets.md) | `hive-prediction-markets` | `prediction_markets` | Research prediction markets, events, outcomes, market stats, traders, holders, and trades. |

## MCP runtime resources and prompts

Resources are the low-cost runtime reference layer agents should read before
guessing from static docs. Prompts are optional workflow starters exposed by
`prompts/list` on the root MCP endpoint.

| Resource URI | Name | Title | MIME type | Purpose |
| --- | --- | --- | --- | --- |
| `hive://providers` | `providers` | Provider Availability | `application/json` | List of all Hive data providers and their current availability |
| `hive://categories` | `categories` | Tool Categories | `application/json` | All Hive tool categories, route paths, and tool counts |
| `hive://tools` | `tools` | Tool Catalog | `application/json` | Lightweight catalog of callable provider tools for this MCP endpoint, including category, provider, and summaries |
| `hive://toolsets` | `toolsets` | Task Toolsets | `application/json` | Canonical task-oriented workflow map for selecting the right Hive tools without hiding raw endpoints |
| `hive://task-canaries` | `task-canaries` | Task Canary Fixtures | `application/json` | Configured task-level canaries, probe tools, stable fixtures, and runtime status semantics |
| `hive://skills` | `skills` | Hive Agent Skills | `application/json` | Agent skill pack metadata for installing Hive, discovering tools, and executing canonical crypto research tasks |
| `hive://status` | `status` | Runtime Status Contract | `application/json` | Provider, task, and observability status metadata using Hive's standard runtime status taxonomy |

| Prompt | Label | Arguments | Purpose |
| --- | --- | --- | --- |
| `analyze_token` | Analyze Token | `token` (required): Token name, symbol, or contract address; `chain` (optional): Blockchain network such as ethereum, solana, or base | Deep analysis of a token using direct market, security, on-chain, holder, and social tools. |
| `compare_protocols` | Compare Protocols | `protocols` (required): Comma-separated list of protocol names | Compare DeFi protocols using direct TVL, fees, yields, and chain-distribution tools. |
| `portfolio_audit` | Portfolio Audit | `address` (required): Wallet address to audit | Audit a wallet using direct balance, position, approval, and transaction tools. |

## SDKs

- **TypeScript MCP client adapter** (official, npm): npm install hive-mcp-client; registry: https://www.npmjs.com/package/hive-mcp-client; adapter source: https://github.com/hive-intel/hive-sdk/tree/main/client — Official typed adapter for custom TypeScript applications, published on npm at v0.1.5 (`npm install hive-mcp-client`). The public source mirror is hive-intel/hive-sdk, with the adapter under client/. REST remains available for non-TypeScript stacks.
- **Python** (community-maintained, pypi): pip install hive-intelligence; registry: https://pypi.org/project/hive-intelligence/ — Community-maintained on PyPI by `utsavhive`; not built from the canonical Hive MCP repo and currently 12+ months behind the public MCP/REST contract. Python users running production workloads should call the REST API directly until the official Python SDK is published.

For languages without a current official package, call the REST API directly using the endpoints above.

## Pricing for agents

- Demo: Free. 10,000 monthly credits, 30 req/min rate limit.
- Analyst: $129 per month. 500,000 monthly credits, 500 req/min rate limit.
- Pro: $499 per month. 2,000,000 monthly credits, 1,000 req/min rate limit.
- Enterprise: Custom. Custom req/min rate limit.

In the normal execution path, one MCP tool call or REST execution request costs one credit regardless of upstream provider or payload size. Authenticated discovery requests can also count against the current backend quota, so cache tool catalogs in production clients. Full plan details at https://www.hiveintelligence.xyz/pricing or in machine-readable form at https://www.hiveintelligence.xyz/pricing.md.

## Safety guidance for agents

- **Always ground transaction-signing on a security check.** Before recommending a user sign anything involving an unknown token, call `get_token_security` and `detect_rugpull`.
- **Respect temporal scope.** When a user asks a historical question, inspect the live schema for time-window, timestamp, block, cursor, or range arguments before execution. If the selected tool only returns current data, say so instead of presenting latest data as historical.
- **Discover tools at runtime.** Do not hard-code provider tool schemas — call `tools/list`, read `hive://tools`, or query authenticated REST `https://mcp.hiveintelligence.xyz/api/v1/tools`. Tool surface evolves.
- **Read error envelopes, do not guess.** Hive returns explicit error structures with provider attribution. Do not synthesize plausible answers when an upstream provider is degraded.

## Identity and disambiguation

Hive Intelligence is an AI data infrastructure company for blockchain and market data. It is distinct from:

- Hive.io (the blockchain network)
- Hive.com (project management software)
- HIVE Digital (the bitcoin mining company)

Confusable names, different entities. The canonical domain is `hiveintelligence.xyz`.

## Contact

- Website: https://www.hiveintelligence.xyz
- Support: support@hiveintelligence.xyz
- Contact form: https://www.hiveintelligence.xyz/contact
- GitHub: https://github.com/hive-intel/hive-sdk
- Twitter / X: https://x.com/Hive_Intel
- Telegram: https://t.me/HiveIntelligence

## Related machine-readable files

- /llms.txt — full site index and key facts
- /llms-full.txt — extended site corpus
- /pricing.md — machine-readable pricing
- /agent-onboarding/SKILL.md — install-path decision tree
- /sitemap.xml — full URL index
- /robots.txt — crawler access policy