CryptoCapi
Use CryptoCapi’s MCP server to query deterministic crypto analysis engines from an agent, with verifiable seals where available; access is pass-gated and the public demo key is limited.
get_insight: analyze one coin bycoin_id(e.g.,bitcoin), choosingview(pulse/alpha) andengine(radar/quant_plus);alphareturns deep analysis withaudit_trailandprotocol_hash. Demo key: BTC/ETH only, always alpha.get_signal: get a Quant Pro quantitative signal for a trading pair such asBTCUSDT; requires the Quant Pro pass.batch_signals: get Quant Plus signals for 1–50 coin identifiers in one call; demo key: BTC/ETH only; unavailable assets returnavailable: false.scan_market: rank market assets by strategy (balanced,aggressive,conservative) with optionallimit1–50; requires the Market Scan pass and uses the curated Quant Plus universe.All tools are read-only, idempotent, non-destructive, open-world MCP tools; missing passes produce explanatory errors.
Provides signed, verifiable crypto market analysis for Bitcoin via CryptoCapi tools such as get_insight, including pulse/alpha views and Quant Plus reproducible signals; batch_signals also supports Bitcoin.
Provides signed, verifiable crypto market analysis for Ethereum via CryptoCapi tools such as get_insight, including pulse/alpha views and Quant Plus reproducible signals; batch_signals also supports Ethereum.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@CryptoCapi give me the signed insight for Bitcoin"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Paste this and it works
No account, no sign-up, no key. It goes in your client's MCP configuration:
{
"mcpServers": {
"cryptocapi": {
"command": "npx",
"args": ["-y", "@cryptocapi/mcp"]
}
}
}That block is the format for Cursor (.cursor/mcp.json in the project, or
~/.cursor/mcp.json for all of them), Windsurf and Claude Desktop. Other
clients use a different shape:
VS Code (GitHub Copilot). The file is .vscode/mcp.json, the key is
servers rather than mcpServers, and each server carries "type": "stdio".
{
"servers": {
"cryptocapi": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@cryptocapi/mcp"]
}
}
}Claude Code and Codex. One command, no file to edit. In Claude Code,
without --scope user it only applies to the project where you ran it.
claude mcp add --transport stdio cryptocapi -- npx -y @cryptocapi/mcp
codex mcp add cryptocapi -- npx -y @cryptocapi/mcpChatGPT, no. Its connectors only reach remote MCP servers, and this one runs locally, over stdio.
With that you can already ask your agent about Bitcoin or Ethereum and get the analysis signed by the two engines that sign it. How far the demo goes is covered below.
Using it with Claude for trading analysis? The guide Trading bot with Claude, without made-up prices (in Spanish) has example prompts and the rules that keep the model from inventing numbers.
Related MCP server: CoinGecko MCP - Cryptocurrency Price & Market Data
Four tools, three engines and Market Scan
Tool | Engine | What it returns | What it requires |
| Radar | Analysis of an asset. The |
|
| Quant Plus | The same asset signed by the deterministic engine, with a | Quant Plus pass |
| Quant Plus | Signals for several assets in one call | Quant Plus pass |
| Quant Pro | Quantitative signal for a trading pair | Quant Pro pass |
| Market Scan (ranking layer) | Market ranking by strategy | Market Scan pass |
There are five rows for four tools because get_insight is the door to two engines, and each one asks for its own pass. Having Radar Alpha does not open Quant Plus through that same tool: change the engine parameter and the required pass changes too.
Market Scan is not a fourth engine. It computes no signals of its own: it ranks the market on top of Quant Plus signals. It is sold separately because it is a different product.
Until 2026-08-30 there were three more tools (get_market_summary, get_prices, get_macro) that returned third-party data: market cap and fear and greed, CoinGecko prices and FRED macro series. They were retired because CryptoCapi is not an aggregator: its engines sign derived intelligence, and third-party data is an internal input. An agent asking "how is the market doing?" would grab the summary and leave with third-party data without touching an engine. Those endpoints still exist in the REST API; what was retired is the agent seeing them as tools.
Each engine is bought separately, so having one does not unlock the others. The descriptions name the engine that is needed, not a generic "PRO", so the agent doesn't waste attempts on tools its key can't open. When it tries anyway, the error tells it which pass is missing and which one it has, instead of a bare 403.
Watch out for a format detail that trips agents up: get_signal takes a trading pair (BTCUSDT) and batch_signals takes coin identifiers (bitcoin). Same engine, two formats, and each field spells it out in its schema.
Try it without signing up
The default configuration uses the public demo key. No account needed.
{
"mcpServers": {
"cryptocapi": {
"command": "npx",
"args": ["-y", "@cryptocapi/mcp"],
"env": { "CRYPTOCAPI_API_KEY": "demo_btc_eth_public" }
}
}
}The env is optional: with no variable at all, the package falls back to the public demo key on its own, so command and args are enough.
What the demo key reaches, measured on 2026-08-30 against the published version:
Engine | Tool | With the demo key |
Radar |
| ✅ bitcoin and ethereum only, always in the |
Quant Plus |
| ✅ bitcoin and ethereum only, reproducible seal with |
Quant Plus |
| ✅ bitcoin and ethereum only, same reach as |
Quant Pro |
| ❌ closed, for any pair |
Market Scan |
| ❌ closed, for any strategy |
The closed ones don't fail because of the coin, they always fail: get_signal with BTCUSDT, Bitcoin's pair, also returns 403. The bitcoin-and-ethereum restriction applies to get_insight and batch_signals, nothing else.
So in a first session, get_insight and batch_signals answer, and they show the product: the signed analysis, on Bitcoin, from the two engines that sign it.
To try all three engines and Market Scan with no coin limit you need the free 14-day trial at https://www.cryptocapi.com. That key opens everything while it lasts.
What sets it apart
The engines' response carries an audit_trail that includes a protocol_hash: a seal of the deterministic computation that produced the analysis. This package forwards those values exactly as they came from the API, without re-serializing them, because reformatting a single number would be enough for the hash to stop verifying.
You don't have to take our word for it: ask your agent for the analysis with get_insight(coin_id="bitcoin", view="alpha") and compare the protocol_hash in that output with the one from the same query made directly against the API.
curl -H "x-api-key: demo_btc_eth_public" \
"https://api.cryptocapi.com/v1/market/insights/bitcoin?view=alpha"They must be identical. If they ever aren't, it's a bug in this package and deserves an issue.
The limit, stated as well: the seal proves the computation is reproducible and that no language model wrote it. It does not prove the analysis is right, and today it is a checksum without a cryptographic signature, so it attests integrity, not origin.
And the package can be verified too
The seal covers the data. That the tarball you download is the one built from this code is covered by something else: it is published from CI with npm provenance, so each version is tied to the commit and the workflow that built it.
npm audit signaturesThe package's npm page also links to the exact commit. It is the same principle as the protocol_hash, applied to the supply chain instead of the data: you don't have to trust us, you can check.
Configuration
Variable | What for | Default |
| Your API key |
|
| API base URL, for development |
|
| Time budget per request, in ms. See below | depends on the tool |
The variables go in the env of the configuration block, or with --env on
the command: in Claude Code, before --transport (if the server name lands
right after --env, it is taken as another variable); in Codex, after the
name.
There isn't a single time budget. Each tool asks for its own, because they
don't cost the same: 20 s for one engine on one asset (get_insight,
get_signal) and 45 s for the ones that go through several (batch_signals,
scan_market). Cutting a market sweep with the budget of a single read gives a
false timeout, and the agent concludes the API is down when it was actually
working.
CRYPTOCAPI_TIMEOUT_MS replaces those budgets, it does not add to them: the
number you set applies to all four tools. Setting it to 30000 to give
scan_market more room actually cuts it, from 45 s to 30. If you change it,
make it higher than the highest one.
When a tool times out, the error states the budget that actually applied, so that is the number to beat.
Get a key with a 14-day trial: cryptocapi.com
Development
npm install
npm run check # types + tests + dependency audit
npm run contrato # checks against the live API that the copied contract still holdsThe tests run on Node's native runner and have no test dependency and never touch the network: the API is faked with node:http. They test the package, not the service, which is what makes them fast and stable.
That runner loads TypeScript without transpiling, so developing requires Node
22.18 or later, even though the published package runs on 20: that is what
engines declares, and what CI checks by starting the built artifact on 20, 22
and 24. It is a development requirement, not a usage one.
That leaves out, on purpose, one half: whether the published package behaves well against the live API and inside an agent. That is what PRUEBAS.md (in Spanish) is for: fourteen manual checks run after every release.
npm run contrato covers part of that half with no human in the loop. The tests
stay off the network on purpose, and that isolation has a blind spot: in
September 2026 four bugs were found and three were about what the API really
sends versus what the package assumed, invisible to a fake server that only
returns what the test wrote. The contract check looks at exactly that, runs
once a day and opens an issue if something moved.
What has been found and not fixed yet lives in PENDIENTES.md (in Spanish), with the reason for each item and enough context to pick it up without re-investigating.
Publishing
It is triggered by a tag and publishes from CI with provenance, to npm and to the MCP Registry:
npm version <patch|minor|major> # also bump the two versions in server.json, and commit
git tag -a vX.Y.Z -m "vX.Y.Z" && git push origin vX.Y.ZThen create the GitHub Release for that tag: Glama only picks up a new version from a Release, not from the tag. Creating it does not publish again, because release.yml runs on tag pushes only.
The workflow first checks that the tag matches the version in package.json and the two in server.json, because an npm version cannot be reused and that mistake can't be undone. Authentication is OIDC trusted publishing, with no token: it is bound to the name of release.yml, so renaming that file breaks publishing.
Once npm is up, a second job sends server.json to the Registry, also over OIDC. It runs separately so that, if the Registry fails, "Re-run failed jobs" retries only that, without going through npm publish again.
For the server, npm version is the single source of the version: it reads the published package.json to declare its serverInfo.version, and a handshake test fails if the two numbers drift apart. The only hand-written literal is the one in server.json, and if it falls behind, the tag doesn't pass the first check.
Versioning
Since 1.0.0 the package follows semver. The contract is the four tools, their parameters and the shape of their responses:
Patch (
1.0.x): fixes and text. Tools and their parameters stay the same.Minor (
1.x.0): new things that break nothing, like one more tool or optional parameter.Major (
2.0.0): removing or renaming a tool, or changing an existing parameter. If it is ever needed, it will be announced first.
Responses are passed through exactly as the API serves them, so a new field can show up in them without a new package version.
License
Available Tools
4 toolsbatch_signalsSeñales de Quant Plus en loteARead-onlyIdempotent
Señales de Quant Plus para varios activos en una sola llamada. Requiere el pase Quant Plus, con una excepción: con la key pública de demostración funciona para bitcoin y ethereum, igual que get_insight. Ojo con el formato: acá van identificadores de moneda ("bitcoin"), no pares de trading, al revés que en get_signal. Solo un universo curado tiene señal, bastante más chico que el conjunto de activos que CryptoCapi sigue: un activo fuera de él vuelve con available: false, y eso es la respuesta correcta, no un error.
Inputs: symbols (required).
get_insightAnálisis de un activoARead-onlyIdempotent
Análisis de un activo con un motor de CryptoCapi: Radar (por defecto) o Quant Plus. La vista "pulse" no exige pase con una key del plan free. La vista "alpha" trae el análisis profundo con el audit_trail y su protocol_hash, que es el sello del cálculo, y requiere el pase Radar Alpha (o Quant Plus si se pide engine="quant_plus"). Con la key pública de demostración, que es la que se usa si no se configura otra, solo responden bitcoin y ethereum, en cualquier vista, y la respuesta es siempre la alpha: otra moneda da 403 aunque se pida "pulse". Al leer la respuesta: z_score mide el MOVIMIENTO del último día cerrado (UTC), no el del día en curso, contra los movimientos pasados, y la posición en Bandas de Bollinger (band_position) mide el NIVEL de precio contra su rango. Son ejes distintos: precio en el tercio superior con z_score negativo no es una contradicción, es un nivel alto después de un día que cerró con una caída fuerte. No las mezcles en una sola frase. market_regime lee ese mismo día: BULLISH_TREND y BEARISH_TREND dicen que subió o bajó más de 2 % o cerró fuera de sus bandas, no que haya una tendencia de varios días. volatility_momentum dice si las bandas se abren (EXPANDING) o se cierran (COMPRESSING): contrastá el texto del análisis contra ese campo y contra band_position. last_close_date es el día UTC que miden todos los números: dos lecturas del mismo día dan los mismos valores, y eso no es un dato viejo. Las lecturas anteriores pueden no traer estos campos. En Radar, analysis.detailed_report lo escribe Python: primero la matemática del último día cerrado y después, por cada nota de prensa de las últimas 48 horas, « informa que…» (una línea de un modelo de lenguaje que pasó la verificación contra la nota: cada cifra y cada nombre propio están en ella) o « titula: «…»» (la línea no pasó y sale el titular tal cual). Cada nota trae su enlace en sources_verified. El sentimiento lo da el z_score, no el modelo. No le atribuyas a un medio más de lo que dice su línea. Las stablecoins se leen por la distancia de su último cierre al dólar, no por el z_score. Si la moneda todavía no tiene análisis, la respuesta es {"status":"success","data":null}: no es un error ni un fallo de la consulta, es que ningún motor calculó nada para ese activo. Decilo así, «no hay análisis disponible», y no completes el hueco por tu cuenta.
Inputs: coin_id (required), view, engine.
get_signalSeñal de Quant ProARead-onlyIdempotent
Señal cuantitativa para un par de trading. Requiere el pase Quant Pro: ningún otro motor lo habilita. Ojo con el formato, que no es el mismo que en las otras herramientas: acá va el par ("BTCUSDT"), no el identificador de moneda ("bitcoin").
Inputs: symbol (required).
scan_marketRanking de Market ScanARead-onlyIdempotent
Ordena el universo curado de Quant Plus por cuánto se movió cada activo el último día cerrado, medido contra su propio umbral de anomalía (threshold_proximity: 1.0 o más es un día anómalo). No es un camino hacia una señal: el z del día siguiente se mide de nuevo, y en la mayoría de los días el primero queda lejos de su umbral. Cada fila trae last_close_date (el día UTC que miden sus números) y el protocol_hash y la engine_version de la lectura de Quant Plus que ordena. Las stablecoins entran solo con la paridad rota (PEG_BROKEN), medidas por su distancia al dólar. Requiere el pase Market Scan, que se vende aparte de Quant Plus aunque el ranking se arme sobre sus señales. El ranking cubre el universo curado de Quant Plus, bastante más chico que el conjunto de activos que CryptoCapi sigue, así que pedir 50 puede devolver bastante menos. Eso no es un fallo.
Inputs: limit, strategy.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool maps to a distinct engine and access tier: get_insight (Radar/Quant Plus analysis per coin), scan_market (Quant Plus ranking), batch_signals (Quant Plus multi-coin), get_signal (Quant Pro trading pair). The main risk is get_signal vs batch_signals, which sound similar, but the descriptions explicitly contrast the input formats ("BTCUSDT" vs "bitcoin") and passes, so an agent can tell them apart.
All four names are snake_case with a verb/noun shape (get_insight, get_signal, scan_market) and one noun-led outlier (batch_signals). The convention is consistent and readable, with only a minor deviation in the batch_ prefix style.
Four tools for a crypto analytics/signal service is compact and each covers a genuinely different surface (single-asset insight, market scan, multi-asset batch, trading-pair signal). It is slightly thin—no dedicated asset-discovery or history tool—but nothing is redundant.
Core analytical workflows are covered, but there is no tool to list supported assets, engines, or historical readings, which the descriptions repeatedly reference. Agents will discover the curated universe only by trial and error via available:false or 403 responses, a notable gap for a data service.
Maintenance
Related MCP Connectors
Pay-per-call DeFi and macro intel for AI agents. x402 USDC tools via streamable HTTP /api/mcp.
Crypto market & research MCP for AI agents: prices, derivatives, options, on-chain, sentiment, news.
Crypto MCP with 21 tools for market data, DeFi, analytics, and sentiment. Post-paid USDC billing.
Real-time crypto & tokenized equity intelligence MCP over x402 USDC on Base mainnet.
Related MCP Servers
- AlicenseBqualityDmaintenanceHigh-performance CCXT MCP server for cryptocurrency exchange integration2494 npm145MIT
- AlicenseNot gradedqualityBmaintenanceOfficial CoinGecko API MCP Server for Crypto Price & Market Data. More details: https://docs.coingecko.com/reference/mcp-server1,147 npm59TypeScriptApache 2.0
- AlicenseNot gradedqualityDmaintenanceAI-powered crypto signal intelligence for 20 assets (BTC, ETH, SOL, etc). 6 scoring dimensions: whale activity, technical analysis, derivatives flow, narrative strength, sentiment, market structure. Market regime detection (TRENDING/RANGING), portfolio optimization, and accuracy tracking. 9 read-only MCP tools. Free via MCP, $0.001 USDC via x402 on Base for REST API.3MIT
- AlicenseNot gradedqualityDmaintenanceCrypto intelligence MCP: 104 tools for market data, ML signals, on-chain analytics, derivatives, and Bittensor subnets. Pay-per-call via x402 USDC on Base/Solana/Algorand/Stellar or $9.99/mo API key.MIT