Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
作者提供的说明(英文原文)
Crypto MCP: 105 tools across news, market, on-chain, sentiment, indicators, agent generation.
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。
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。
选择你使用的 AI 应用,然后复制设置。
BCA_API_KEY, BCA_API_BASE标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
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 …”。
打开 Claude Desktop → Settings → Developer → Edit Config,把下面的内容粘贴到 claude_desktop_config.json
{
"mcpServers": {
"blockchainacademics": {
"command": "npx",
"args": [
"-y",
"@blockchainacademics/mcp"
],
"env": {
"BCA_API_KEY": "your-value",
"BCA_API_BASE": "your-value"
}
}
}
}在终端中运行这条命令
claude mcp add blockchainacademics --env BCA_API_KEY=your-value --env BCA_API_BASE=your-value -- npx -y @blockchainacademics/mcp运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"blockchainacademics": {
"command": "npx",
"args": [
"-y",
"@blockchainacademics/mcp"
],
"env": {
"BCA_API_KEY": "your-value",
"BCA_API_BASE": "your-value"
}
}
}
}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。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"blockchainacademics": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@blockchainacademics/mcp"
],
"env": {
"BCA_API_KEY": "your-value",
"BCA_API_BASE": "your-value"
}
}
}
}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。
这个服务器在你的电脑上运行,而 ChatGPT 只能连接互联网上的服务器。请选择其他应用。
你的选择通过 Cookie 保存在这台设备上,无需账户。
BCA_API_BASE env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。
它在你的电脑上运行;安装后,工具列表会显示在你的 AI 应用中。
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7日
注册名称已通过 GitHub 账户 blockchainacademics 验证,该账户拥有此仓库
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
它在你的电脑上运行,没有互联网端点
在你的电脑上运行,改用本地设置
它在你的电脑上运行;我们只能读取互联网端点的工具列表
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。
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。
你需要填写:BCA_API_KEY
它在你的电脑上运行。你的 AI 应用会在需要时启动它,数据不会经过作者的服务器(除非它自己调用在线服务)。
GitHub 仓库最后更新于–,最新版本(0.6.3)于5天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。你的电脑上还需要有 Node.js。
同一分类中有:Spec Driven Development, Remote Desktop Commander, Agent Data Tools: Utilities
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
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
103 tools: JSON repair, CSV, hashes, cron, dates, units, validators. 25 free calls.
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。
The canonical crypto MCP server for AI agents. 105 tools spanning 3,800+ editorial articles, 150+ entity dossiers, 38 academy lessons in 8 courses, aggregated market + on-chain data, proprietary indicators, and async agent-backed generation, all exposed as MCP tools your AI agent can call natively.
| Category | Count | Examples |
|---|---|---|
| Content & corpus | 7 | search_news, get_article, get_entity, get_explainer, list_entity_mentions, get_as_of_snapshot |
| Market data | 4 | get_price, get_ohlc, get_market_overview, get_pair_data |
| On-chain | 4 | get_wallet_profile, get_tx, get_token_holders, get_defi_protocol |
| Sentiment + crypto Twitter | 6 | get_sentiment, get_social_pulse, get_fear_greed, get_ct_pulse, get_x_signals, get_kol_consensus |
| Proprietary indicators | 6 | get_coverage_index, get_narrative_strength, get_sentiment_velocity, get_editorial_premium, get_kol_influence, get_risk_score |
| Agent-backed (async, Pro+) | 6 | generate_due_diligence, generate_tokenomics_model, summarize_whitepaper, translate_contract, monitor_keyword, get_agent_job |
| Directories | 16 | list_stablecoins, list_yields, list_aggregators, list_mcps, list_vcs, list_jobs, get_job, search_grants, get_grant, build_custom_indicator, … |
| Fundamentals | 6 | get_tokenomics, get_audit_reports, get_team_info, compare_protocols, check_rugpull_risk |
| Chain-specific | 4 | get_solana_ecosystem, get_l2_comparison, get_bitcoin_l2_status, get_ton_ecosystem |
| Markets microstructure | 5 | get_funding_rates, get_options_flow, get_liquidation_heatmap, get_exchange_flows, predict_listing |
| Narrative / meta | 5 | track_narrative, get_ai_agent_tokens, get_depin_projects, get_rwa_tokens, get_prediction_markets |
| Regulatory | 4 | get_regulatory_status, track_sec_filings, get_mica_status, get_tax_rules |
| Security | 4 | check_exploit_history, check_phishing_domain, get_bug_bounty_programs, scan_contract |
| Memes / degen | 4 | track_pumpfun, track_bonkfun, check_memecoin_risk, get_degen_leaderboard |
| Services (revenue) | 3 | book_kol_campaign, request_custom_research, submit_listing |
| History time-series | 4 | get_history_prices, get_history_sentiment, get_history_correlation, get_history_coverage |
| Compute / AI crypto | 2 | get_compute_pricing, get_ai_crypto_metrics |
| Corpus meta | 7 | list_entities, get_topic, search_academy, get_trending, get_unified_feed, list_sources, get_recent_stories |
| Memos + theses + social | 6 | list_memos, get_memo, list_theses, get_thesis, get_social_signals, get_social_signals_detail |
| Currencies | 2 | list_currencies, get_currency_feed |
Every tool response carries attribution.citations[] (each entry a cite_url, an as_of time and a source_hash) and a meta.status that says how complete the answer is (see Response status).
Tools whose data source is not connected yet (bonk.fun, CoinGlass liquidations and exchange flows, the degen leaderboard, the memecoin analyzer, among others) answer with meta.status: "unseeded" and a reason instead of failing, and their descriptions say so.
LLMs hallucinate about crypto. BCA ships ground-truth editorial content with full attribution. Plug this MCP server into Claude Desktop, LangChain, Eliza, or any MCP-compatible agent and your model queries the BCA corpus like any other tool, with citations, timestamps, and source hashes on every response.
The package ships with a public demo key baked in. No signup needed to try it. Add to claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": { "blockchainacademics": { "command": "npx", "args": ["-y", "@blockchainacademics/mcp"] } } }
Restart Claude Desktop. 17 tools work without a key, with real data and real citations:
get_price, get_trending, get_fear_greed, get_market_overview, search_news, get_article, get_sentiment, get_entity, get_explainer, get_recent_stories, get_topic, search_grants, get_grant, list_jobs, get_job, search_academy, get_as_of_snapshot
get_entity works by slug without a key; its ticker form needs one. Every other tool answers with a clean upgrade message (BCA_TIER_LOCKED) that links to the free signup. Demo calls share a daily cap: 100 calls a day across all demo users, and 20 a day per IP.
Get a free key at blockchainacademics.com/crypto-api/signup (2,000 calls/month per user) and add the env block:
{
"mcpServers": { "blockchainacademics": { "command": "npx", "args": ["-y", "@blockchainacademics/mcp"], "env": { "BCA_API_KEY": "bca_live_xxxxxxxxxxxxxxxx" } } } }
Restart Claude Desktop. All 105 tools appear in the tool picker: a free key opens 52 of them, and paid plans open the rest (see API Key).
npm install @blockchainacademics/mcp
import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";
const transport = new StdioClientTransport({ command: "npx", args: ["-y", "@blockchainacademics/mcp"], env: { BCA_API_KEY: process.env.BCA_API_KEY! }, }); const client = new Client({ name: "my-agent", version: "0.0.1" }, { capabilities: {} }); await client.connect(transport);
const res = await client.callTool({ name: "search_news", arguments: { query: "stablecoin regulation", limit: 5 }, }); console.log(res.content);
See examples/research-agent.ts for a full tool-chaining loop.
search_newsFull-text search across 3,800+ editorial crypto articles.
| arg | type | required | description |
|---|---|---|---|
query | string (1-512) | yes | Search query |
entity | string | no | Entity slug filter (e.g. ethereum) |
ticker | string | no | Ticker filter (e.g. ETH) |
since | ISO 8601 | no | Earliest publish date |
topic | string | no | Topic filter (e.g. regulation) |
limit | number (1-50) | no | Default 10 |
offset | number | no | Results to skip, for paging (default 0) |
Example: { "query": "circle IPO", "since": "2026-01-01T00:00:00Z", "limit": 5 }
get_entityFetch an entity dossier. Provide exactly one of:
| arg | type | description |
|---|---|---|
slug | string | e.g. "vitalik-buterin", "circle"; works without a key |
ticker | string | e.g. "ETH", "SOL" (case-insensitive); returns every matching entity and needs an API key |
Aliases in slug form resolve to the canonical entity (cz → changpeng-zhao, maker → makerdao, bsc → bnb-chain, …).
get_explainerFetch BCA Academy material (38 lessons in 8 courses). Provide exactly one of:
| arg | type | description |
|---|---|---|
slug | string | Course slug (e.g. "defi-fundamentals", "understanding-bitcoin") for the course and its lesson index, or lesson slug (e.g. "decentralized-exchanges") for the lesson with its full text |
topic | string | Topic keyword that resolves to matching academy material, e.g. "defi" |
meta.status says how complete an answer is. Branch on it before you trust data:
meta.status | Meaning | Where the detail is |
|---|---|---|
complete | A full, fresh answer. | no meta.diagnostic |
partial | Data is returned, but some inputs or sections are missing or defaulted, or one upstream source failed. | meta.diagnostic.reason, plus meta.diagnostic.missing (what is absent) or, for a failed source, vendor and upstream_status |
stale | Data is returned, but it is older than the tool promises. | meta.diagnostic.reason, meta.diagnostic.as_of (when the data was produced) and meta.diagnostic.age_seconds |
unseeded | There is no real data yet: nothing matched, the table is empty, or the source is not connected. data may still carry the echoed query or placeholders. | meta.diagnostic.reason and meta.diagnostic.eta |
Lists that page by offset (search_news, search_grants, list_entities, list_entity_mentions, search_academy) set meta.pageInfo.hasNextPage when more results exist.
Every response includes a structured attribution block with a citations array:
{
"data": { ... }, "attribution": { "citations": [ { "cite_url": "https://www.coingecko.com/en/coins/bitcoin", "as_of": "2026-11-01T12:34:56Z", "source_hash": "sha256:..." }, { "cite_url": "https://blockchainacademics.com/...?src=...", "as_of": "2026-11-01T12:34:56Z", "source_hash": "sha256:..." } ] }, "meta": { "status": "complete", "request_id": "...", "pageInfo": { ... } } }
citations[0] is the primary source: the upstream page when the data has one (the original article, CoinGecko, Etherscan, DefiLlama, a regulator), otherwise the BCA page for the record, otherwise the tool's entry in the API reference. A BCA link follows the upstream one only where the site has a page for the data. When your agent surfaces BCA content to a user, attribute via citations[0].cite_url. Fields are preserved as null when upstream omits them so downstream agents can detect missing provenance.
Not required to try it. When BCA_API_KEY is unset, the server falls back to a baked-in public demo key: the 17 tools listed under Install work, and every other tool answers with a clean upgrade message.
For more, get a free key at blockchainacademics.com/crypto-api/signup (2,000 calls/month): it opens the 52 free-tier tools. Paid tiers unlock the rest, such as agent-backed research generation, crypto Twitter intelligence and the proprietary indicators; each tool's description names its tier when it is not free.
Set BCA_API_KEY in your MCP client env. Optionally override BCA_API_BASE (default https://api.blockchainacademics.com). BCA_API_BASE_URL is also accepted for backward compatibility.
Every failed tool call returns an MCP tool result with isError: true and a JSON body:
{ "error": { "code": "BCA_AUTH", "message": "..." } }
A request that is not a valid MCP call at all (arguments that are not an object, for example) is answered by the MCP SDK with a JSON-RPC error instead.
| Code | Meaning |
|---|---|
BCA_AUTH | Invalid BCA_API_KEY (HTTP 401/403), or a key that cannot be sent as a header value (a line break inside it); the key itself is never repeated in the message |
BCA_TIER_LOCKED | Tool not in your current tier's allowlist (HTTP 403). Without a key this applies to 88 of the 105 tools; the message includes the signup URL. |
BCA_RATE_LIMIT | Rate limit exceeded (HTTP 429; honor Retry-After) |
BCA_UPSTREAM | BCA API returned an error status or malformed JSON. For a 4xx the message carries the API's own detail (up to 500 characters) inside an fence; a 5xx stays generic. |
BCA_NETWORK | Network failure or 20s timeout exceeded |
BCA_BAD_REQUEST | Invalid tool arguments, caught before any request is sent. The message lists each problem as field: problem, e.g. Invalid arguments for search_news: limit: Number must be less than or equal to 50. |
npm install
npm run build # tsc -> dist/ npm test # node:test smoke suite npm run dev # tsx src/index.ts (stdio)
This repo uses [`pre-co
只显示了开头部分;其余内容请在仓库中阅读