# Sharpe Terminal (npm · @sharpe-terminal/mcp-server)

Live crypto derivatives data: funding rates, open interest, basis, arbitrage and listings

- Trust score: 65/100 (medium)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- remote · `www.sharpe.ai`: 39/100, [markdown](https://verifymcp.io/servers/ai-sharpe-mcp/api-mcp.md), [page](https://verifymcp.io/servers/ai-sharpe-mcp/api-mcp)
- npm · `@sharpe-terminal/mcp-server`: 65/100 (this document), [markdown](https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server.md), [page](https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server)

## Channel facts

- Registry: `npm`
- Package: `@sharpe-terminal/mcp-server`
- Version: `4.3.0`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Supply Chain Security**: 98/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - No install/post-install scripts declared.
  - 31 of 92 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 19/100
  - Repository check failed: no source repository is declared.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 1 days ago).
  - Security-disclosure policy not yet verified: we couldn't inspect the source repository.
- **Schema Quality & AI Usability**: 72/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 13439 tokens (~268/item across 50 items; 45 tools + 5 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 17/100
  - Stability observed for 5 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (93% of tools); any adoption earns full credit.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "get_arbitrage_cex_spot_transfer" implies "transfer" and declares readOnlyHint instead, contradicting what its own name says it does.
  - An AI judge read all 47 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the Sharpe Terminal MCP server?

Sharpe Terminal runs locally as an npm package, launched with npx -y @sharpe-terminal/mcp-server. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add ai-sharpe-mcp -- npx -y @sharpe-terminal/mcp-server
```

### Cursor

```json
{
  "mcpServers": {
    "ai-sharpe-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@sharpe-terminal/mcp-server"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "ai-sharpe-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@sharpe-terminal/mcp-server"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add ai-sharpe-mcp -- npx -y @sharpe-terminal/mcp-server
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "ai-sharpe-mcp": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "@sharpe-terminal/mcp-server"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add ai-sharpe-mcp --command npx --arg -y --arg @sharpe-terminal/mcp-server
```

### Hermes

```yaml
mcp_servers:
  ai-sharpe-mcp:
    command: "npx"
    args: ["-y", "@sharpe-terminal/mcp-server"]
```

### Netclaw

```json
{
  "McpServers": {
    "ai-sharpe-mcp": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "@sharpe-terminal/mcp-server"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add ai-sharpe-mcp -t stdio -c npx -a -y @sharpe-terminal/mcp-server
```

### Other

```json
{
  "mcpServers": {
    "ai-sharpe-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@sharpe-terminal/mcp-server"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-21 (score 65, +1)

No change was recorded against any check on this day. Stability & Change Management went from 13 to 17. That category is still filling its 30-day observation window: 4 days of observed history at the previous scan, 5 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-20 (score 64, +10)

- [security improvement] Known CVEs: unverified → pass
- [functional improvement] Dependency health: unverified → 0.84

### 2026-11-19 (score 54, −10)

- [security regression] Known CVEs: pass → unverified
- [functional regression] Dependency health: 0.85 → unverified
- [functional] Package version: 4.1.0 → 4.3.0
- [functional] Package version: 4.1.0 → 4.2.0

### 2026-11-18 (score 64, +1)

- [functional regression] Tool coverage: 100% → 95%
- [functional improvement] Stability: unverified → 0.07
- [functional] Package version: 4.0.1 → 4.1.0

### 2026-11-16 (score 63)

First indexed and scored.

## MCP tools (45)

### `get_api_coverage` (~38 tokens)

API data coverage

Get public metadata about what data is available across all Sharpe Terminal products, including supported exchanges, coins, timeframes, and chart types.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_borrow_rates` (~268 tokens)

Margin borrow rates

Get spot-margin borrow rates per coin across Binance, OKX, Bybit and Gate.io (the venues that publish them without an account), for pricing a short spot leg (reverse cash-and-carry: sell spot, long perp) or a leveraged long. Each row is one coin with every venue's quote cheapest first: hourlyRate (fraction per hour) and annualRate (hourlyRate × 8,760, simple, fraction per year: 0.05 = 5%), the source endpoint and its time, plus lowest/highest venue. A venue that does not price a coin is absent, never 0; Gate.io covers its 100 most-traded loan currencies. `venues` says which venues answered (ok / stale / failed / unsupported). KuCoin, Bitget, MEXC, BingX and HTX publish no keyless rate. Filter with coin and exchanges.

Input parameters:

- `coin` (string): Borrowed asset, case-insensitive (BTC, USDT, PEPE).
- `exchanges` (string): Comma-separated venues, case-insensitive: Binance, OKX, Bybit, Gate.io (KuCoin, Bitget, MEXC, BingX and HTX are accepted and reported unsupported). Default: all.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_cex_spot_transfer` (~518 tokens)

CEX spot-transfer arbitrage scanner

Get live CEX spot-transfer arbitrage routes (buy on one exchange, withdraw, deposit and sell on another) after exact network matching (ambiguous network names refused), token-contract matching, withdrawal/deposit status and minimums, withdrawal fees, per-venue taker fees (buyTakerFee, sellTakerFee), depth and slippage. netReturn is the net per trip as a fraction of the notional (0.001 is 0.1%), like every spread, fee and slippage figure; minNet filters it as a fraction. Routes netting above zero come first (executionStatus executable); up to 50 near-miss routes follow as indicative with indicativeReason below_fees (or stale_quote / stale_network). Books are judged on 60 s, networks on 10 minutes. transferConfirmations is the sell venue's deposit confirmation count (transferEta is its deprecated label); meta.transferScan counts coins scanned, venues answering, routes evaluated and profitable, and the best net (bestNetReturn).

Input parameters:

- `coin` (string): Optional base coin ticker, for example BTC, ETH, or SOL.
- `cursor` (string): Opaque cursor from the previous page.
- `exchanges` (string): Comma-separated exchange filter, for example Binance,Bybit,OKX.
- `limit` (integer): Maximum number of rows returned.
- `minDepthUsd` (number): Minimum executable depth in USD when available.
- `minNet` (number): Minimum of the board's net figure, as a fraction (0.05 is 5%): `netApr` on the dated basis board, `netRollApy` on the calendar board, `netCarryApr` (the fee-adjusted annualized net carry, not the edg…
- `notional` (number): Position notional in USD. On the dated-futures basis, calendar-spread and perp-dated carry scanners a row is executable only when its book depth (level 1 on the stored snapshot; on a live read a top-…

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_cross_exchange` (~593 tokens)

Cross-exchange arbitrage

Get cross-exchange funding arbitrage across every funding venue and asset class (crypto plus RWA perps: equity, commodity, FX, index — pairs never cross classes; lot spellings such as 1000PEPE and kPEPE pair as one asset), one row per coin (its best executable pair, else its best no_depth pair, else its largest gross differential), ranked with executable rows first by netApr, then no_depth rows, then the other indicative rows, by netApr or gross apr. Executable = book-priced spread, legs within 5%, fresh quotes, top-of-book depth reported by both venues and ≥ $10,000; otherwise indicativeReason says why (no_depth: a venue reports no size; reference_only, price_gap, thin_book, oi_below_notional, stale_quote, unpriced). netApr nets both venues' taker fees and the adverse entry spread (a favourable spread is never credited) and is present only on book-priced rows; unknown spread/OI/volume are null, never zero. Rows carry pairStatus, breakEvenDays, roundTripFeeRate, the legs' contract symbols, fundingAsOf and per-leg price times; limit / cursor page the board.

Input parameters:

- `assetClass` (string): Underlying filter: "all" (default), "crypto", or "rwa" (equity, commodity, FX and index perps).
- `cursor` (string): Opaque cursor from the previous page's `pagination.cursor` (100 rows a page when `limit` is omitted).
- `exchanges` (string): Comma-separated list of venues to compare. Omit for every funding venue in the funding tracker's registry that the instrument master does not mark not live (`X-Data-Venues` lists them).
- `limit` (integer): Return at most this many rows, the first in board order (executable rows first by netApr, then `no_depth` rows, then the rest, each by netApr or apr). Default: every row. On v1 a limited answer carri…
- `minOiUsd` (number): Minimum open interest in USD on the smaller leg. Default 0 (no floor). Legs with unknown OI fail a positive floor, and so does a leg on a venue left out of the API's open-interest totals (MEXC).
- `minVolUsd` (number): Minimum 24h perp volume in USD on the smaller leg. Default 0 (no floor). Legs with unknown volume fail a positive floor.
- `sector` (string): Narrative filter: only crypto coins in that narrative's CoinGecko category (the constituents /api/narratives/data lists). Omit or "all" for every coin. A coin the category does not list, and every RW…

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_dated_futures_basis` (~546 tokens)

Dated futures basis scanner

Get cash-and-carry arbitrage rows for buying spot and selling dated futures (Binance, OKX, Bybit), ranked by net annualized basis after fees (netApr: basis at spot ask / futures bid less 2 spot + 2 futures taker fills, × 365 / days to expiry recomputed per request; a fraction per year, 0.05 is 5%, and minNet filters it as one). Rows at least 3 days out rank first (nearExpiry rows last); executable rows first, then snapshot_quote rows (executable on a live read), then the rest, each with indicativeReason (stale_snapshot, no_depth, depth_below_notional, depth_beyond_top_of_book, oi_below_notional, no_volume, backwardation_needs_borrow, snapshot_quote). basis, annualizedBasis and fees (the round-trip fee at each venue's standard rate) are fractions too.

Input parameters:

- `coin` (string): Optional base coin ticker, for example BTC, ETH, or SOL.
- `cursor` (string): Opaque cursor from the previous page.
- `exchanges` (string): Comma-separated exchange filter, for example Binance,Bybit,OKX.
- `limit` (integer): Maximum number of rows returned.
- `marginType` (string): Futures margin filter.
- `minDepthUsd` (number): Minimum executable depth in USD when available.
- `minNet` (number): Minimum of the board's net figure, as a fraction (0.05 is 5%): `netApr` on the dated basis board, `netRollApy` on the calendar board, `netCarryApr` (the fee-adjusted annualized net carry, not the edg…
- `minOiUsd` (number): Minimum open interest in USD.
- `minVolumeUsd` (number): Minimum 24h futures or spot volume in USD.
- `notional` (number): Position notional in USD. On the dated-futures basis, calendar-spread and perp-dated carry scanners a row is executable only when its book depth (level 1 on the stored snapshot; on a live read a top-…

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_fixed_floating_funding` (~339 tokens)

Fixed vs floating funding (Pendle Boros)

Compare Pendle Boros fixed funding rates with realised floating funding. For every live Boros market (a venue's perp to a maturity: Binance, OKX, Bybit, Gate, KuCoin, Hyperliquid incl. xyz, Lighter) returns impliedApr (Boros's fixed rate), floatingApr (the venue's current funding as Boros reports it), realised7dApr and realised30dApr (Sharpe's stored funding for the same contract, annualised as acc × 365 / days), and spread7dApr / spread30dApr = implied − realised. A positive spread means locking the fixed rate (sell yield on Boros, receive fixed) beats what the floating leg has paid; negative means the fixed rate is cheap to pay. All rates are fractions per year (0.05 = 5%), gross of fees. Realised figures are null with realised7dNullReason / realised30dNullReason when the contract is not covered or its history is incomplete, never 0. Also serves bestBidApr / bestAskApr, daysToMaturity, openInterestUsd (one side) and a Boros trade link. Filter with coin and venue.

Input parameters:

- `coin` (string): Underlying asset, case-insensitive (BTC, ETH, HYPE, BRENTOIL).
- `venue` (string): Filter to one venue, case-insensitive: Boros's name (Binance, OKX, Hyperliquid, Gate, Kucoin, Lighter, Bybit) or our funding store's (Gate.io, KuCoin, tradeXYZ).

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_funding_spread_history` (~573 tokens)

Funding spread history and carry backtest

Get a pair's historical funding spread and carry backtest from stored settlements. mode=spot-perp takes venue (+ optional contract) and direction (short = buy spot, short perp; long = sell spot, long perp, charged the venue's current borrow APR where published: borrowApr, borrowCostIncluded); mode=cross-exchange takes longVenue and shortVenue (+ optional longContract / shortContract). Legs are aligned by settlement time with each contract's real interval (never an assumed 8h). Returns series (spreadApr and legAprs annualized, cashflow and cumulativeCarry as fractions of one leg's notional) and backtests for 1/7/14/30/60 days and the requested days (1-90): fundingCollected, fees (each venue's entry+exit taker fees), netReturn, grossApr, netApr, negativeCarryDays, maxDrawdown, breakEvenDays, coverageRatio, complete and estimatedSettlements (settlements that are still pre-settlement estimates; series points carry estimated). Fractions: multiply by 100 for percent.

Input parameters:

- `coin` (string, required): Base coin, for example BTC.
- `contract` (string): Spot-perp: the venue's perp contract symbol, for example BTCUSDT. Optional: without it the venue's contract for the coin with the largest current open interest is used.
- `days` (integer): Backtest window in days (1-90): the series covers it, and `backtests` lists it after the shorter standard windows.
- `direction` (string): Spot-perp: `short` buys spot and shorts the perp (collects positive funding); `long` sells spot and longs the perp (the venue's current borrow APR is charged where it publishes one: `borrowCostInclud…
- `longContract` (string): Cross-exchange: the long perp's contract symbol. Optional, resolved as `contract` is.
- `longVenue` (string): Cross-exchange: the venue of the long perp.
- `mode` (string): `spot-perp`: one venue's spot and perp (needs `venue`). `cross-exchange`: a long perp and a short perp (needs `longVenue` and `shortVenue`).
- `shortContract` (string): Cross-exchange: the short perp's contract symbol. Optional, resolved as `contract` is.
- `shortVenue` (string): Cross-exchange: the venue of the short perp.
- `venue` (string): Spot-perp venue: Binance, OKX, Bybit, Bitget, Gate.io, KuCoin, MEXC, BingX or HTX. Required in the default `spot-perp` mode: a call without it answers 400.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_futures_calendar_spread` (~536 tokens)

Futures calendar spread scanner

Get near-versus-far dated futures calendar spread rows (adjacent pairs plus monthly/quarterly pairs at least 7 days apart) with executable forward yield (forwardYield), the mark-or-mid reference forward yield that sets curve and direction (referenceForwardYield, priceReference), and lock-in net roll APY after 4 futures + 2 spot taker fills (netRollApy, fees). Yields and APYs are fractions per year (0.05 is 5%), fees and nearBasis / farBasis fractions; minNet filters netRollApy as a fraction. Pairs with a near leg under 3 days or a gap under 7 days are belowTenorFloor and rank last; within the rest, executable pairs that clear their fees come first, then snapshot_quote pairs, then the rest. Rows carry nearDays, gapDays and indicativeReason when not executable.

Input parameters:

- `coin` (string): Optional base coin ticker, for example BTC, ETH, or SOL.
- `cursor` (string): Opaque cursor from the previous page.
- `exchanges` (string): Comma-separated exchange filter, for example Binance,Bybit,OKX.
- `limit` (integer): Maximum number of rows returned.
- `marginType` (string): Futures margin filter.
- `minDepthUsd` (number): Minimum executable depth in USD when available.
- `minNet` (number): Minimum of the board's net figure, as a fraction (0.05 is 5%): `netApr` on the dated basis board, `netRollApy` on the calendar board, `netCarryApr` (the fee-adjusted annualized net carry, not the edg…
- `minOiUsd` (number): Minimum open interest in USD.
- `minVolumeUsd` (number): Minimum 24h futures or spot volume in USD.
- `notional` (number): Position notional in USD. On the dated-futures basis, calendar-spread and perp-dated carry scanners a row is executable only when its book depth (level 1 on the stored snapshot; on a live read a top-…

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_perp_dated_carry` (~591 tokens)

Perp versus dated futures carry scanner

Get perp funding versus dated futures carry rows: the dated contract's basis against the PERP (not spot) at the chosen trade's executable sides (datedBasis), against the funding the perp is expected to pay to expiry (expectedFundingToExpiry: the 7-day realised average over the settlements left, the current print under 3 days; fundingProjection says which), ranked by fee-adjusted annualized net carry (netCarryApr, which minNet filters as a fraction). Every rate is a fraction: APRs per year (0.05 is 5%); the basis, edges and expected funding to expiry, and fees, of the notional. Rows carry direction (Long perp / Short dated or Short perp / Long dated), crossVenue (the legs are usually on two venues: separate accounts and collateral), perpSymbol (the perp contract), perpFundingApr, funding7dApr, funding7dCumulative, fundingIntervalHours, fundingSettlementsToExpiry, datedLegPrice / perpLegPrice, fees, nearExpiry (under 3 days: ranked last) and indicativeReason when not executable (snapshot_quote = executable only on a live read).

Input parameters:

- `coin` (string): Optional base coin ticker, for example BTC, ETH, or SOL.
- `cursor` (string): Opaque cursor from the previous page.
- `exchanges` (string): Comma-separated exchange filter, for example Binance,Bybit,OKX.
- `limit` (integer): Maximum number of rows returned.
- `marginType` (string): Futures margin filter.
- `minDepthUsd` (number): Minimum executable depth in USD when available.
- `minNet` (number): Minimum of the board's net figure, as a fraction (0.05 is 5%): `netApr` on the dated basis board, `netRollApy` on the calendar board, `netCarryApr` (the fee-adjusted annualized net carry, not the edg…
- `minOiUsd` (number): Minimum open interest in USD.
- `minVolumeUsd` (number): Minimum 24h futures or spot volume in USD.
- `notional` (number): Position notional in USD. On the dated-futures basis, calendar-spread and perp-dated carry scanners a row is executable only when its book depth (level 1 on the stored snapshot; on a live read a top-…

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_arbitrage_spot_perp` (~464 tokens)

Spot-perp arbitrage

Get spot-perpetual funding-capture rows with live funding, verified spot or margin availability, diagnostic basis, gross APR, and net APR after the venue's taker fees, the adverse entry basis and a known borrow cost, in the board's order (executable rows first by net APR, then no_spot_depth rows, then the rest, by net APR or gross APR). No row is executable today: spot prices are last trades with no book, so a row passing every other check is indicative no_spot_depth. Each row carries perpSymbol, perpOpenInterestUsd, fundingIsStale, executionStatus with indicativeReason (no_spot_depth, unpriced, price_gap, identity_mismatch, stale_quote, borrow_unpriced, borrow_limit, thin_book, oi_below_notional), pairStatus, breakEvenDays, requiresBorrow / borrowCostIncluded, apr7d / carrySignFlip7d (the 7-day carry against the current direction), fundingAsOf and per-leg price times. minOiUsd floors the perp leg's open interest (unknown fails a positive floor); limit / cursor page the board (the MCP tools ask for a 25-row page unless given either: the whole board is ~2.4 MB).

Input parameters:

- `cursor` (string): Opaque cursor from the previous page's `pagination.cursor` (100 rows a page when `limit` is omitted).
- `direction` (string): Filter by perpetual-leg direction: short for positive funding capture, long for negative funding capture.
- `exchange` (string): Filter by exchange name, or "all" for all exchanges.
- `limit` (integer): Rows per page (1-1000), in board order. Omitted with no `cursor`: every row. On v1 a paged answer carries `pagination`.
- `minOiUsd` (number): Minimum open interest in USD on the perpetual leg (`perpOpenInterestUsd`). Default 0 (no floor). A row whose open interest is unknown fails a positive floor, and so does a row on a venue left out of…

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_correlation_matrix` (~136 tokens)

Correlation matrix

Get a price correlation matrix for crypto and supported TradFi assets over a specified calendar period. Useful for portfolio analysis and diversification.

Input parameters:

- `ids` (string): Comma-separated asset ids: CoinGecko ids for crypto (e.g. bitcoin) and short ids for TradFi (e.g. sp500, gold). Maximum 10. If omitted, returns the default set. Ids no source recognises are listed in…
- `period` (string): Lookback period for correlation calculation.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_derivatives_overview` (~58 tokens)

Market-wide derivatives overview

Get a derivatives market overview including aggregated open interest, 8-hour-equivalent funding rates, top coins by open interest (each top_coins_oi entry is an asset, its lot spellings included), and exchange/coin counts.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_ecosystems` (~147 tokens)

Ecosystems data

Get ecosystem-level market data (Ethereum, Solana, BNB, Arbitrum, Base, etc.) including market cap, volume, TVL, and performance metrics.

Input parameters:

- `correlation` (string): Set to "true" to include correlation data.
- `ecosystem` (string): Ecosystem to return detail for: the config id (e.g. bnb) or the URL slug (e.g. bnb-chain, ethereum, solana). An unknown value returns 400 invalid_parameter listing the valid values.
- `excludeNative` (string): Set to "true" to exclude the native token from aggregated metrics.
- `timeframe` (string): Timeframe for correlation data.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_exchange_listings` (~124 tokens)

New listings hub data

Get the New Listings hub payload: weekly + monthly aggregated counts by narrative and exchange, plus the last-90-day recent listings. Optionally filter by narrative slug or single exchange. Counts only resolved spot listings (CoinGecko id + mapped narrative); new tokens awaiting resolution are missing, so use get_listing_events for a complete count of the newest listings.

Input parameters:

- `exchange` (string): Filter to a single exchange.
- `narrative` (string): Filter to a single narrative slug (e.g. `ai-agents`, `memes`, `layer-1`).

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_funding_rates` (~1512 tokens)

Funding rates

Get perpetual funding rates across every supported venue (see sharpe://exchanges for the current list). Returns current snapshots, accumulated rates over multiple windows, or per-coin historical rates. Units: `rate` is a decimal fraction settled over that row's own `interval_hours` (1h, 2h, 4h, 8h or 24h depending on venue and market). It is not a percent and never implicitly 8h. On most venues `rate` is the in-progress estimate for the coming settlement, not what was last charged; `rate_kind` says which (`estimate`, `last_settled` on Kraken/Crypto.com/Bitfinex, `trailing_8h` on Deribit), and type=history serves the settled rates. Compare or rank venues on the served rate_8h (rate × 8 / interval_hours) or apr (rate × 8760 / interval_hours), both null when the row has no usable interval. Accumulated rows carry realized sums (acc_1d/acc_7d/acc_30d/acc_90d/acc_1y) instead of a per-interval rate, each annualised over the history behind it (acc_apr_1d … acc_apr_1y), and per window a `coverage` entry (coverage_ratio, complete, listed_within_window, halted_within_window): compare acc_* across contracts only on complete windows where listed_within_window and halted_within_window are both false. Filter with coin, exchange, margin (linear/inverse) and asset_class; asset_class is unfiltered by default, so tokenized-equity, commodity, index and FX perps are included alongside crypto. `coin` names an asset: PEPE returns every PEPE contract, the 1000PEPE and KPEPE lot contracts included. Current and accumulated rows carry asset_id, asset_symbol, lot_multiplier (asset units per contract unit, e.g. 1000 on 1000PEPEUSDT), instrument_status (the venue's own status) and is_live (null = unknown). Page the full book with limit and cursor, or order it by open interest with sort. For type=current the API also answers with aggregates instead of rows: extremes (the highest and lowest markets by rate_8h above an optional min_oi_usd floor), stats (the whole book's mean and OI-weighted fundin…

Input parameters:

- `asset_class` (string): Narrow the response to one underlying category. The funding book is not crypto-only. Tokenized equity, commodity, index and FX perps trade on the same venues, so this is unfiltered by default. Applie…
- `coin` (string): The coin to return: a ticker (e.g. BTC), a lot or venue spelling (e.g. `1000PEPE`, `kPEPE`), or the full name (e.g. `bitcoin`, resolved to BTC). Case-insensitive, one coin per request. `coin` names a…
- `cursor` (string): Opaque pagination cursor from a previous response.
- `days` (integer): Number of days of history to retrieve (type=history only).
- `exchange` (string): Filter to a single venue. Accepts the slug (e.g. `gate-io`), the display name as returned in the `exchange` field (e.g. `Gate.io`), or the ccxt id. The venue registry is not frozen, so read the `shar…
- `extremes` (integer): type=current only: answer with the N highest and N lowest rows by `rate_8h` (`extremes.highest`, `extremes.lowest`) among fresh rows with a usable interval, and how many rows each check left out (`ex…
- `limit` (integer): Maximum number of records to return per page.
- `margin` (string): Narrow the response to one collateral convention: `linear` for quote-margined contracts, `inverse` for coin-margined ones. Unfiltered by default. Applies to type=current and type=accumulated; ignored…
- `min_oi_usd` (number): type=current, with `extremes` only: leave out rows whose open interest in USD is below this floor. A row with unknown open interest fails a floor above 0.
- `sort` (string): type=current only: order the rows by `open_interest`, ascending, or descending with the leading `-`. Rows with unknown open interest come last either way. Applied before `limit`/`cursor`. Default: th…
- `stats` (string): type=current only: `1` answers with the whole book's funding (`market`: mean and open-interest-weighted APR, the count of positive and negative rates, and how many of the tracked venues have a fresh…
- `summary` (string): type=current or type=accumulated, with `coin` only: `1` answers with the coin's funding across venues (`summary`) instead of the rows. For type=current: the average and median 8-hour rate, how they w…
- `type` (string): Type of funding rate data to retrieve.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_funding_settlement` (~773 tokens)

Funding fee settlement

Get the dollar value of funding actually paid at each settlement -- not just the rate -- and which side, longs or shorts, paid it. Every response carries both the per-coin row breakdown and a totals block (net/long_paid/short_paid) for the requested window and class, so 'who paid funding today' is read directly from totals rather than summed from a page of rows; totals always reflects the full requested class regardless of any coin/exchange filter on the rows. window=current (the default) is the next 24 hours of funding at current rates -- each contract's forward rate x 24 / its own interval hours, priced live off open interest -- always estimated, never realised. 1d/3d/7d are realised windows; one whose coverage is below 90% carries partial: true (complete: false) because its history is still accruing, not a completed total, and halted_within_window says a contract stopped settling inside the window (listed_within_window is always null: no listing time is known). Sign convention: a window's net = long_paid - short_paid; positive means longs paid more than they received (a cost to longs), negative means shorts paid more. Both long_paid and short_paid are always non-negative magnitudes. Missing is null, never 0. class scopes to one category as on the funding board (crypto, equity, commodity, fx, index), or rwa = every non-crypto class; the five classes' totals, and crypto + rwa, always equal the all totals.

Input parameters:

- `class` (string): Asset-class scope, the same categories as the funding board: `all`, `crypto`, `equity`, `commodity`, `fx` or `index` (one `asset_class` each). `rwa` is every non-crypto class at once (equity + commod…
- `coin` (string): Comma-separated coins, for example BTC,ETH. Each names an asset as on /funding/rates: full names resolve (bitcoin -> BTC) and any spelling finds the asset's one row (PEPE, 1000PEPE and KPEPE all answ…
- `exchange` (string): Filter to coins settling on a single venue. Accepts the slug, the display name as returned in the `venues[].exchange` field, or the ccxt id. The venue registry is not frozen, so read the `sharpe://ex…
- `expand` (string): Set to `venues` to include each row's per-venue breakdown. Omitted by default to keep the default payload small.
- `limit` (integer): Maximum number of rows returned.
- `offset` (integer): Zero-based row offset for pagination.
- `sort` (string): Sort field for the selected `window`, optionally prefixed with "-" for descending. Rows with no value for the sort field (an unpriced coin) always sort last, regardless of direction.
- `window` (string): Settlement window. `current` is the next 24 hours of funding at current rates (each contract's forward rate x 24 / its own interval hours, priced live off open interest) and labelled `estimated`. `1d…

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_futures_coins` (~53 tokens)

Available futures coins

Get up to 1,000 coins with futures data, including funding, open-interest, liquidation, long/short, CVD, and dated-basis capability flags plus observed venue coverage and market-cap rank.

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_futures_data` (~417 tokens)

Futures chart data

Get futures chart data including open interest, volume, funding rate history, liquidations, long/short ratios, basis, term structure, and OHLCV. OI rows sum every contract a venue stored that hour; a row with `oi_contracts` holds another number of contracts than the venue's newest hour (`oi_contracts_latest`): fewer is partial, and a change across two counts is coverage, not positions. Time series are oldest-first; when one exceeds the response cap its newest rows are kept (truncated_fields marks kept: latest). Without limit, oi-snapshot returns each exchange's latest reading; pass limit/cursor to page raw rows oldest-first. chart=cvd returns candle rows with `cvd_delta_usd` and the running `cvd_usd` (CVD = running sum of (2 * taker_buy_volume - volume) * close; only Binance candles carry taker volume). Liquidation quantity is already USD notional (never multiply by price). Composite charts return their raw input legs tagged `_type`, plus server-computed values: funding-cost settlement rows carry `funding_paid_usd` (USD the settlement moved, + = longs paid; absent when no OI prices it), crowding-risk adds per-bucket Binance price rows with `crowding_score` (−100 crowded shorts … +100 crowded longs; null when a leg is missing), and oi-snapshot rows carry `oi_share` (0-1 share of the listed venues' latest OI, each venue's latest row). returns-* return raw 1h candles.

Input parameters:

- `chart` (string, required): Futures chart identifier.
- `coin` (string): Base coin ticker.
- `cursor` (string): Opaque pagination cursor from a previous response.
- `exchanges` (string): Comma-separated list of exchanges to include (e.g. Binance,Bybit).
- `limit` (integer): Maximum number of data points per page.
- `timeframe` (string): Lookback window for the data.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_gem_finder` (~297 tokens)

Gem finder

Get market-ranked token rows, optionally scoped to one chain (chain=solana, base, ...), with market cap, volume, price and market-cap change, circulating supply, ATH and ATL history, nullable FDV/MCap, the home chain (chainSlug, what chain filters on) plus every supported chain the asset is deployed on including bridged copies (chainSlugs), observed exchange availability, quote freshness, and enrichment coverage. Delisted or halted markets whose upstream quote is over 24 hours old are excluded rather than reporting a weeks-old price change as current. Ambiguous same-ticker assets do not receive symbol-only exchange attribution. No wallet signal or proprietary score is returned.

Input parameters:

- `chain` (string): Keep only assets whose home chain (chainSlug) is this chain. A token bridged to the chain (chainSlugs) does not match: Chainlink is an Ethereum asset, not a Solana one. A chain's own coin is homed on…
- `cursor` (string): Opaque, snapshot-bound pagination cursor from a previous response. Cursors expire when the cache refreshes; restart without a cursor after an expiry error.
- `limit` (integer): Maximum number of tokens to return per page.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_global_overview` (~373 tokens)

Global market overview

Get the whole-market derivatives board in one call: total open interest with its crypto-versus-RWA split (the headline includes RWA perps: equities, commodities, ETFs, indices, FX), 24h liquidations on Binance, Bybit and Deribit only (a lower bound, not the market total: OKX, Hyperliquid, Gate, Bitget and other venues are not in it), average Wilder RSI(14) across the top 100 crypto perps by market cap, the Altcoin Season Index (share of the top 50 by market cap beating BTC over 90 days, excluding stablecoins and wrapped/staked derivatives), open interest broken down by asset class, and a ranked per-asset row table. The tab parameter selects the universe: crypto, equity, preipo, etf, index, commodity or fx. Rows carry price, 24h return, funding rate (interval-true, never assumed 8h), 24h volume, market cap, open interest with 1h and 24h changes (returns and changes are fractions: 0.01 is 1%), plus liquidations on crypto or a venue count on RWA tabs. Ranks are by open interest descending. Liquidations are crypto-only because no RWA venue publishes a liquidation feed. Use this for 'what is the market doing' questions; use get_funding_rates for per-venue crypto funding detail and get_rwa_perp_funding for RWA carry and cross-venue spreads.

Input parameters:

- `cursor` (string): Opaque pagination cursor from a previous response.
- `limit` (integer): Maximum rows per page (1-1000). Enables cursor pagination.
- `tab` (string): Universe to return rows for. The metric block is identical on every tab.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_heatmap` (~601 tokens)

Market heatmap

Get the market heatmap: coins (top10, top20, top30 or top50 by market cap, or one narrative's or ecosystem's members) or one tile per narrative or ecosystem, with price return (returns: 1h, 4h, 24h, 7d, 30d, 1y), funding APR (avgFundingApr), RSI(14), RVOL, open interest and OI value change (USD, oiChange24h); returns, funding and OI change are fractions (0.054 is 5.4%). Only eligible coins count: at least $1M market cap and $100K plausible volume, CoinGecko data under 48 hours old, and no stablecoins, wrapped or staked copies, LP tokens, funds or tokenized stocks and ETFs. Tiles are weighted baskets of every eligible member, with member coverage in `constituents` and the median member in `median`. Pass `ex` (CoinGecko ids, e.g. bitcoin,ethereum) to leave coins out; aggregate tiles are rebuilt without them. Refreshed every 3 hours, stale after 6 hours; candle and funding times are in `as_of`.

Input parameters:

- `category` (string): Coins-mode universe. `top10`, `top20`, `top30` and `top50` (the default) are the N largest eligible coins by market cap; the legacy `top-100` is an alias of `top50`. Otherwise one of the 15 narrative…
- `ex` (string): Comma-separated CoinGecko coin ids to leave out, lower case (e.g. `bitcoin,ethereum`). Coins mode: removed after the Top N cut and never refilled, so `top10` with two member ids answers 8 coins. Aggr…
- `mode` (string): Heatmap grouping mode.

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_insider_selling` (~292 tokens)

Insider selling pressure

Get insider selling pressure signals across crypto assets. Scores require cross-venue negative funding and combine OI, basis, CVD (basis and CVD exist for futures-covered majors only), sell-side flow, and supporting wallet/supply context; oi_change_24h, return_24h and the other shares and changes are fractions (0.054 = 5.4%); read score_interpretation: a score missing funding, price or derivatives inputs is a conservative lower bound, while missing on-chain inputs are re-weighted and listed in incomplete_signal_groups. coin returns one coin's row as the scorer wrote it, dated by its own updated_at (score null and stale once it stopped qualifying).

Input parameters:

- `coin` (string): One coin by its CoinGecko id (e.g. pepe): its row exactly as the scorer wrote it, whatever its score, and the answer is dated by that row's own `updated_at`. A coin that stopped qualifying answers `s…
- `limit` (integer): Maximum number of coins to return.
- `min_score` (number): Filter to coins scoring at or above this value.

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_listing_events` (~234 tokens)

Canonical listing events

Get canonical listing lifecycle events across CEXs and Perp DEXs, including source URLs, confidence, source-observation freshness (for official overlays, the newest listing the venue reports), partial status, and per-provider coverage. Pass the opaque cursor through unchanged.

Input parameters:

- `asset_class` (string): Filter by listed instrument class.
- `confidence` (string): Filter by source confidence.
- `cursor` (string): Opaque two-stream cursor returned by the previous page. Pass it through unchanged.
- `days` (integer): Lookback window in days.
- `event_type` (string): Filter by listing lifecycle event type.
- `exchange` (string): Filter to a single exchange or perp DEX.
- `from` (string): ISO date or datetime lower bound.
- `limit` (integer): Max rows returned.
- `market_type` (string): Filter by market type.
- `narrative` (string): Filter by Sharpe narrative slug.
- `status` (string): Filter by event status.
- `to` (string): ISO date or datetime upper bound.
- `venue_type` (string): Filter by venue type.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_listing_exchanges` (~83 tokens)

New listings exchange registry

Get the typed listing connector registry, including supported exchanges, venue types, market coverage, source documentation URLs, enabled status, polling group, and parser confidence.

Input parameters:

- `enabled` (string): Filter to enabled or disabled connectors.
- `market_type` (string): Filter exchange coverage by supported market type.
- `venue_type` (string): Filter exchange coverage by venue type.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_market_overview` (~91 tokens)

Market overview

Get BTC and ETH price and 24h return (return_24h), total crypto market cap and 24h volume, BTC dominance (btc_market_share), the market cap's 24h return (market_cap_return_24h), and the Fear & Greed Index. Returns and shares are fractions (0.573 is 57.3%). Auxiliary source fields may be null when unavailable.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_memecoin_launches` (~167 tokens)

Memecoin launches

Get recently launched tokens from recognized memecoin launchpads, screened by age, liquidity, volume, and transactions. Generic new DEX pairs are excluded; response includes complete/partial chain coverage.

Input parameters:

- `chains` (string): Comma-separated chain IDs. Default `solana,bsc`: the chains with a recognised meme launchpad (Pump.fun on Solana; Four.meme and Flap on BNB Chain). Other chains are scanned but return no rows until a…
- `days` (integer): Launch lookback window in days, from 1 to 30.
- `limit` (integer): Maximum launch rows to return, from 1 to 200.
- `profile` (string): Screening profile for liquidity, volume, and transaction thresholds.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_memecoins` (~139 tokens)

Memecoins data

Get memecoin narrative data across grouped categories (Dog, Cat, Frog, AI, Solana, Base, Pump.fun, etc.) with market cap, market share, volume, performance, and freshness fields.

Input parameters:

- `coinHistory` (string): Include per-coin price history where retained. Enum: 24h, 7d, 1m, 1y.
- `historical` (string): Include historical time-series data for the specified window.
- `narrative` (string): Memecoin narrative slug to filter by (e.g. dog-coins, cat-coins, frog-coins, ai-memes, trump-coins).

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_mindshare` (~105 tokens)

Mindshare data

Get narrative mindshare (each narrative's share of CoinGecko watchlist users, not social-media posts) rankings, token-level watchlist-share rows, rolling windows, or historical mindshare snapshots.

Input parameters:

- `historical` (string): Set to true for historical snapshots.
- `narrative` (string): Optional narrative slug.
- `timeframe` (string): Historical timeframe.
- `tokens` (string): Set to true for token rows.
- `window` (string): Rolling snapshot window.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_narrative_bench` (~224 tokens)

Narrative Bench

Rank every crypto narrative's index (AI Agents, Layer 1, DeFi, Memes, RWA, ...) against BTC buy-and-hold over a window (7D, 30D, 90D, 180D, YTD, 1Y, 3Y, 5Y; default 90D): each narrative's return, growth over BTC ((1 + return) / (1 + BTC return) − 1), the prior window's, max drawdown, its constituents and weights, and the coins its trim set aside. Each index holds the narrative's 15 largest coins over $10M, none above 10%, rebalanced monthly, with each month's most extreme movers set aside so no single coin carries it. Daily closes.

Input parameters:

- `timeframe` (string): The window, ending at the latest close: 7D to 180D count days; 1Y, 3Y and 5Y start on the same calendar date years earlier; YTD starts at the previous year's last close.

### `get_narratives` (~262 tokens)

Narratives data

Get narrative-level market data (L1, L2, DeFi, AI Agents, DePIN, RWA, Gaming, etc.) including market cap, volume, performance, and social metrics. sort orders the list by return24h and top/bottom add the N highest and lowest narratives by return24h (a narrative without one is never ranked).

Input parameters:

- `bottom` (integer): List only: add `bottom`, the N narratives with the lowest 24h change.
- `correlation` (string): Set to "true" to include correlation data.
- `narrative` (string): Narrative to return detail for: the config id (e.g. dex) or the URL slug (e.g. dex-tokens, layer-1, ai-agents). An unknown value returns 400 invalid_parameter listing the valid values.
- `sort` (string): List only: order `narratives` by 24h return, highest (`-return24h`) or lowest (`return24h`) first; a narrative without a 24h return is last either way.
- `timeframe` (string): Timeframe for correlation data.
- `top` (integer): List only: add `top`, the N narratives with the highest 24h change.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_news` (~127 tokens)

News feed

Get aggregated crypto news articles from multiple sources, with optional title search and filtering by category or coin. Articles are deduplicated so only the best source per story is shown.

Input parameters:

- `category` (string): Filter by news category.
- `coin` (string): Filter by coin ticker.
- `limit` (integer): Maximum number of articles to return.
- `offset` (integer): Number of articles to skip for pagination.
- `q` (string): Search article titles across the stored news corpus.
- `since` (string): ISO 8601 timestamp. Only return articles published after this time.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_open_interest_market` (~170 tokens)

Open interest across the market

Get the whole derivatives market's open interest hour by hour over up to 30 days (default 30): each venue's crypto and RWA open interest, each asset class's total (crypto, equities, commodities, ETFs, pre-IPO, indices, FX), the largest assets of each class with their split by venue, every asset over $100,000 now, and BTC's hourly price. One side in US dollars; MEXC, BingX and any venue with no reading at the last hour (Deribit, 8 hours late) left out, listed in left_out; tradeXYZ counted as Hyperliquid. For one coin's open interest by venue use get_futures_data.

Input parameters:

- `days` (integer): Days of hourly series, ending at the last whole hour.

### `get_options_open_interest` (~184 tokens)

Options open interest

Get a coin's options open interest on Deribit, OKX, Bybit and Binance (hourly): totals per venue, by expiry with put/call ratio and max pain, by strike (one expiry or all), each venue's hourly history, and the options-to-futures open interest ratio. US dollars, one side; Deribit's reading is 8 hours older (it publishes open interest 8 hours late). History starts 2026-11-17.

Input parameters:

- `coin` (string): Base coin ticker (BTC, ETH, SOL, XRP, ...).
- `days` (integer): Days of hourly `history` and `futures_ratio`.
- `exchange` (string): One venue; every venue when omitted.
- `expiry` (string): UTC date (YYYY-MM-DD) whose options `strikes` lists; every expiry when omitted.

### `get_price_prediction` (~87 tokens)

Price prediction

Get deterministic directional scores and sub-signal breakdowns for a coin. Optional 7D/30D values are volatility-scaled heuristic scenarios, not calibrated targets or probabilities.

Input parameters:

- `coin` (string): Ticker, canonical coin slug, or CoinGecko ID to get predictions for (for example, `BTC` or `bitcoin`). A curated coin's ticker resolves to its slug.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_pump_dump` (~286 tokens)

Pump and dump detection

Get pump-and-dump manipulation risk signals across crypto assets. Scores combine derivatives pressure, holder concentration, liquidity quality, DEX flow, and price behavior; read score_interpretation: a score missing funding, price or derivatives inputs is a conservative lower bound, while missing on-chain inputs are re-weighted and listed in incomplete_signal_groups. oi_change_24h, return_24h and the other shares and changes are fractions (0.054 = 5.4%). coin returns one coin's row as the scorer wrote it, dated by its own updated_at (score null and stale once it stopped qualifying).

Input parameters:

- `coin` (string): One coin by its CoinGecko id (e.g. pepe): its row exactly as the scorer wrote it, whatever its score, and the answer is dated by that row's own `updated_at`. A coin that stopped qualifying answers `s…
- `limit` (integer): Maximum number of coins to return.
- `min_score` (number): Filter to coins scoring at or above this value.
- `phase` (string): Filter to one manipulation phase label.

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_recent_listings` (~268 tokens)

Recent new listings

Get a flat feed of recent token listings. Defaults remain backward-compatible with the legacy spot listing feed, which holds only resolved spot listings (CoinGecko id + mapped narrative), so the newest tokens and perps are missing and `limit` truncates it without a cursor; pass an event filter or `from` for the complete event feed. Event filters add source-observation freshness, partial status, and provider coverage so agents can reject incomplete opportunity scans.

Input parameters:

- `asset_class` (string): Optional listed instrument class filter.
- `confidence` (string): Optional source confidence filter.
- `cursor` (string): Cursor returned by previous page.
- `days` (integer): How many days back to include. Default 90.
- `event_type` (string): Optional event-feed lifecycle event filter.
- `exchange` (string): Filter to a single exchange.
- `from` (string): ISO date or datetime lower bound.
- `limit` (integer): Max rows returned.
- `market_type` (string): Optional event-feed market type filter.
- `narrative` (string): Filter to a single narrative slug.
- `status` (string): Optional event status filter.
- `to` (string): ISO date or datetime upper bound.
- `venue_type` (string): Optional event-feed venue type filter.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_rug_check_security` (~180 tokens)

Rug Check token security

Get rug-check token contract security and liquidity risk signals, and the Rug Check score the /rug-check page shows, computed by the API: score.value (0-100, higher is safer; null only when no pillar had a scorable figure), score.verdict, score.data_completeness and score.engine_version, with the checks[] behind it.

Input parameters:

- `address` (string, required): Token contract address.
- `chainId` (integer, required): EVM chain ID (1 Ethereum, 56 BNB Smart Chain, 8453 Base, 42161 Arbitrum, 137 Polygon, 10 Optimism, 43114 Avalanche, 146 Sonic, 80094 Berachain, and 29 more), or 1399811149 for Solana and 728126428 fo…

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_rug_check_trending` (~32 tokens)

Rug Check trending tokens

Get trending tokens suitable for rug-check review.

Input parameters:

- `limit` (integer): Maximum tokens to return.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_rwa_perp_funding` (~455 tokens)

RWA perp funding rates

Get funding and carry for real-world-asset perpetuals (RWA perps: stocks, pre-IPO, ETFs, indices, commodities, FX) across 13 venues: CEXs (Binance, Bybit, OKX, Bitget, Gate, Kraken xStocks), perp DEXs (Aster, Lighter, Extended, GRVT, ApeX, Pacifica) and Hyperliquid HIP-3 builder markets. Rates are interval-true and mechanism-true: funding_apr annualizes each venue's own settlement interval (never an assumed 8h), and a both-sides rollover is reported separately as borrow_apr_annual (a cost, never mixed into funding spreads). Marks more than 2% off the cross-venue median are flagged is_price_suspect and excluded from stats and spreads. type=history returns stored session-tagged settlements (market/after_hours/weekend/holiday) for Hyperliquid and Kraken only, since 2026-07-28; type=stats returns best carry (trailing-24h realised), top cross-venue spread (comparable funding only), weekend premium, and open interest. funding_status separates a real 0 from a venue's off-hours pause or a market that never funds.

Input parameters:

- `cursor` (string): Opaque pagination cursor from a previous response.
- `days` (integer): Days of settlement history to retrieve (type=history only).
- `limit` (integer): Maximum number of records to return per page.
- `symbol` (string): Underlying ticker (e.g. TSLA, XAU, US500), case-insensitive. Required when type=history. GOOG and GOOGL are distinct share classes.
- `type` (string): Type of rwa-perp data to retrieve.
- `venue` (string): Filter to a single venue, case-insensitive (e.g. Binance, Bybit, OKX, Kraken, Aster, Lighter, Pacifica, or the Hyperliquid HIP-3 builder venue: Hyperliquid). The venue list is registry-driven and gro…

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_stablecoins` (~98 tokens)

Stablecoins data

Get stablecoin overview, detail, or yield data including nominal supply at peg, marked value, nullable peg metrics, mechanism mix, chain supply, velocity coverage, freshness, and APY risk flags. Shares, supply changes, peg deviation and APYs are fractions (0.054 = 5.4%).

Input parameters:

- `slug` (string): Stablecoin slug. Required when type=detail.
- `type` (string): Payload type.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `get_web_traffic` (~128 tokens)

Web traffic data

Get CoinGecko watchlist rankings and snapshots, or market-level signals, for coins and narratives, plus CoinGecko trust-score rankings for exchanges. Measures no site visits or web traffic.

Input parameters:

- `entities` (string): Comma-separated entity IDs for snapshots.
- `mode` (string): Payload mode.
- `sub` (string): Market dataset to return. Required when mode=market; ignored by other modes.
- `tf` (string): Requested history lookback for snapshots. The route clamps to available history when the requested range is unavailable.
- `type` (string): Entity type.

Output parameters:

- `data`
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `search_market_cap` (~64 tokens)

Market cap search

Search crypto and supported TradFi assets by name or ticker. Returns matching assets with asset class, market cap, and nullable price, day return, FDV, ATH, and estimated ATH market cap fields.

Input parameters:

- `q` (string, required): Search query (2-100 characters).

Output parameters:

- `data` (object)
- `meta` (object): Which route answered, what the API said about the data's freshness, and whether the client cut the answer to fit the agent's context window.
- `pagination` (object): Present when the answer is one page: pass `cursor` back to read the next one.

### `analyze_coin` (~58 tokens)

Analyze Coin

Single-coin analysis in one view: funding rates, futures OI, and a deterministic directional score. Requires an API key (SHARPE_API_KEY).

Input parameters:

- `coin` (string, required): Coin ticker (e.g. BTC, ETH, SOL)

Output parameters:

- `coin` (string): Coin the report covers.
- `error` (string): Error message when the call fails.
- `report` (string): Formatted markdown report.

### `market_briefing` (~36 tokens)

Market Briefing

Market briefing with market overview, top narratives, and a funding rate summary. Requires an API key (SHARPE_API_KEY).

Output parameters:

- `coin` (string): Coin the report covers.
- `error` (string): Error message when the call fails.
- `report` (string): Formatted markdown report.

### `find_opportunities` (~44 tokens)

Find Opportunities

Scan all exchanges for the best funding rate arbitrage opportunities across spot-perp basis trades and cross-exchange funding arb. Requires an API key (SHARPE_API_KEY).

Output parameters:

- `coin` (string): Coin the report covers.
- `error` (string): Error message when the call fails.
- `report` (string): Formatted markdown report.

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server#diagnostics

## Score history

- 2026-11-21: 65
- 2026-11-20: 64
- 2026-11-19: 54
- 2026-11-18: 64
- 2026-11-17: 63
- 2026-11-16: 63

## Common questions

### What is the Sharpe Terminal MCP server?

Sharpe Terminal is an MCP server listed in the public MCP registry as ai.sharpe/mcp. Live crypto derivatives data: funding rates, open interest, basis, arbitrage and listings. This page covers its npm package (@sharpe-terminal/mcp-server).

### Is the Sharpe Terminal MCP server safe to use?

Sharpe Terminal scores 65 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the Sharpe Terminal MCP server expose?

Sharpe Terminal exposes 45 tools: get_api_coverage, get_arbitrage_borrow_rates, get_arbitrage_cex_spot_transfer, get_arbitrage_cross_exchange, get_arbitrage_dated_futures_basis, and 40 more. Their descriptions and schemas cost roughly 12,771 tokens of context every time the server is loaded.

### Is the Sharpe Terminal MCP server still maintained?

Sharpe Terminal is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the Sharpe Terminal MCP server under?

Sharpe Terminal declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

### What did VerifyMCP test on the Sharpe Terminal MCP server?

VerifyMCP ran 15 checks on Sharpe Terminal, covering Malware scan, Known CVEs, Install scripts, Dependency health and 11 more. One item could not be completed, a gap on our side rather than a finding about Sharpe Terminal: Security disclosure. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- npm package: https://www.npmjs.com/package/@sharpe-terminal/mcp-server
- Socket report: https://socket.dev/npm/package/@sharpe-terminal/mcp-server
- Website: https://www.sharpe.ai/docs/mcp-server
- Changelog RSS feed: https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server.xml
- Changelog JSON feed: https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server.json
- HTML version of this page: https://verifymcp.io/servers/ai-sharpe-mcp/sharpe-terminal-mcp-server
