# CandleFeed (pypi · candlefeed-mcp)

Rebuild Binance USD-M order book days for spread and depth; candles, funding, OI and liquidations

- Trust score: 59/100 (low)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-10-18

## Components

- pypi · `candlefeed-mcp`: 59/100 (this document), [markdown](https://verifymcp.io/servers/ai-candlefeed-candlefeed-mcp/candlefeed-mcp.md), [page](https://verifymcp.io/servers/ai-candlefeed-candlefeed-mcp/candlefeed-mcp)

## Channel facts

- Registry: `pypi`
- Package: `candlefeed-mcp`
- Version: `0.1.5`
- 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 2026-10-18.

- **Supply Chain Security**: 100/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - Runs hatchling.build at install time, a recognised build step with no custom scripting around it.
  - 2 of 38 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 3 days ago).
  - Security-disclosure policy not yet verified: we couldn't inspect the source repository.
- **Schema Quality & AI Usability**: 54/100
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 1784 tokens (~178/item across 10 items; 10 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 0/100
  - Stability not yet verified: not enough scan history yet (needs a 30-day window).
- **Tool Coverage**: 71/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 0% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 10 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 11 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

**Unverified: 1 category.** A category scored 0 because we could not verify it: a data source with nothing on this package, evidence we could not reach, or a check we could not run. We only credit what we can confirm.

## Install

### How do I install the CandleFeed MCP server?

CandleFeed runs locally as a PyPI package, launched with uvx candlefeed-mcp. 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-candlefeed-candlefeed-mcp -- uvx candlefeed-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "ai-candlefeed-candlefeed-mcp": {
      "command": "uvx",
      "args": [
        "candlefeed-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "ai-candlefeed-candlefeed-mcp": {
      "command": "uvx",
      "args": [
        "candlefeed-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add ai-candlefeed-candlefeed-mcp -- uvx candlefeed-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "ai-candlefeed-candlefeed-mcp": {
      "type": "local",
      "command": [
        "uvx",
        "candlefeed-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add ai-candlefeed-candlefeed-mcp --command uvx --arg candlefeed-mcp
```

### Hermes

```yaml
mcp_servers:
  ai-candlefeed-candlefeed-mcp:
    command: "uvx"
    args: ["candlefeed-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "ai-candlefeed-candlefeed-mcp": {
      "Transport": "stdio",
      "Command": "uvx",
      "Arguments": [
        "candlefeed-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add ai-candlefeed-candlefeed-mcp -t stdio -c uvx -a candlefeed-mcp
```

### Other

```json
{
  "mcpServers": {
    "ai-candlefeed-candlefeed-mcp": {
      "command": "uvx",
      "args": [
        "candlefeed-mcp"
      ]
    }
  }
}
```

## 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.

### 2026-10-15 (score 59, +25)

- [security improvement] Injection markers: unverified → pass
- [security] First check of Judged manipulation: pass
- [security] Stability: Stability not yet verified: not enough scan history yet (needs a 30-day window).
- [functional improvement] Tool coverage: unverified → 100
- [functional improvement] MCP protocol: unverified → pass
- [functional] First check of Tool coverage: 0
- [functional] First check of Schema quality: fail
- [functional] First check of Destructive annotations: pass
- [functional] First check of Schema quality: fail
- [functional] First check of Schema quality: good
- [functional] First check of Tool coverage: 100

### 2026-10-14 (score 34, 0)

- [security] Stability: Stability not yet verified: this server's registry entry declares environment variables (an API key or similar), and its server did not start in our sandbox, so we have no schema to compare.
- [security] Tool safety: Tool safety not yet verified: this server's registry entry declares environment variables (an API key or similar), and its server did not start in our sandbox, so we have no tool text to scan.
- [functional] Schema quality: Schema quality not yet verified: this server's registry entry declares environment variables (an API key or similar), and its server did not start in our sandbox, so we have no schema to assess.
- [functional] Capabilities: Protocol version not yet verified: this server's registry entry declares environment variables (an API key or similar), and its server did not start in our sandbox, so we never saw its MCP handshake.
- [functional] Tool coverage: Tool coverage not yet verified: this server's registry entry declares environment variables (an API key or similar), and its server did not start in our sandbox, so we have no tool definitions to assess.
- [functional] Package version: 0.1.4 → 0.1.5

### 2026-10-12 (score 34)

First indexed and scored.

## MCP tools (10)

### `l2_coverage` (~65 tokens)

Which order book ("book") and tick-trade ("trades") days are published, per symbol: first and
last day, day count, and unpublished days in between with the reason. Public, no key needed.

Input parameters:

- `dataset`
- `symbol`

Output parameters:

- `result` (object)

### `l2_gaps` (~72 tokens)

The public gap log: every stretch on a published day that none of the three capture nodes
recorded, with times, size and the reason. Dates are YYYY-MM-DD. Public, no key needed.

Input parameters:

- `dataset`
- `end`
- `start`
- `symbol`

Output parameters:

- `result` (object)

### `l2_download_day` (~135 tokens)

Download one UTC day of Binance USD-M L2 files (dataset "book": 24 hourly diff files plus the
day's REST snapshots; "trades": tick trades) into the local cache, checking every file's size and
SHA-256. Files already cached with a matching hash are skipped. The 1st of each month is free on
every plan; other days need Pro. Uses CANDLEFEED_API_KEY; without it, only the public no-account sample
days can be downloaded.

Input parameters:

- `dataset` (string)
- `date` (string, required)
- `symbol` (string, required)

Output parameters:

- `result` (object)

### `l2_book_at` (~109 tokens)

Rebuilt order book at a moment (ISO 8601 UTC, e.g. 2026-09-01T13:05:00Z): top `levels` bids and
asks as [price, qty], best bid/ask, mid, spread in bps, and the snapshot it was anchored on.
The day must be downloaded first with l2_download_day.

Input parameters:

- `levels` (integer)
- `symbol` (string, required)
- `time` (string, required)

Output parameters:

- `result` (object)

### `l2_spread_summary` (~128 tokens)

Bid-ask spread statistics for a downloaded day (or an ISO start/end window inside it), sampled
every `freq` (100ms to 5min): mean, median, p95 and max spread in bps, the mid range, and the
biggest 1-minute mid move. series_points > 0 adds an evenly thinned series (max 500 points).

Input parameters:

- `date` (string, required)
- `end`
- `freq` (string)
- `series_points` (integer)
- `start`
- `symbol` (string, required)

Output parameters:

- `result` (object)

### `l2_depth_summary` (~139 tokens)

Resting size within each `bps` band of the mid (default [10, 50]) at a moment, per side, in the
base asset and in USDT, each band flagged complete or not. window_minutes > 0 adds mean/min/max of
that depth over complete one-second samples, with their count. Depth outside the anchor snapshot's
known price window is partial: report completeness per band and exclude incomplete samples from
full-depth statistics. The day must be downloaded first.

Input parameters:

- `bps`
- `symbol` (string, required)
- `time` (string, required)
- `window_minutes` (integer)

Output parameters:

- `result` (object)

### `candles` (~195 tokens)

OHLCV candles (intervals 1m 5m 15m 1h 4h 1d). Free keys get BTC, ETH, SOL, XRP and DOGE on
Binance for the last 30 days; other venues need Advanced. Max 5,000 rows per call.
Rows are oldest first. Without start you get the newest `limit` rows up to a fixed end (end, or
now, sent to the API) from a window that is `limit` intervals long, starting on a bucket boundary (whole buckets within 30 days on Free); with start, the first `limit`
rows from start. The result's `window` and `note` say which range was used.

Input parameters:

- `end`
- `exchange` (string)
- `interval` (string)
- `limit` (integer)
- `start`
- `symbol` (string, required)

Output parameters:

- `result` (object)

### `funding_rates` (~165 tokens)

Funding rate settlements per exchange (funding_rate, mark_price). Free on the five free
symbols for 30 days. Max 5,000 rows per call.
Rows are oldest first. Without start you get the newest `limit` rows up to a fixed end (end, or
now, sent to the API) from a window that is `limit` x 8 h long (x 1 h on Hyperliquid, clipped to 30 days on Free); with
start, the first `limit` rows from start. The result's `window` and `note` say which range was used.

Input parameters:

- `end`
- `exchange` (string)
- `limit` (integer)
- `start`
- `symbol` (string, required)

Output parameters:

- `result` (object)

### `open_interest` (~157 tokens)

Open interest in contracts and USD (intervals 5m 15m 1h 4h 1d). Builder plan and up.
Max 5,000 rows per call.
Rows are oldest first. Without start you get the newest `limit` rows up to a fixed end (end, or
now, sent to the API) from a window that is `limit` intervals long; with start, the first `limit` rows from start. The
result's `window` and `note` say which range was used.

Input parameters:

- `end`
- `exchange` (string)
- `interval` (string)
- `limit` (integer)
- `start`
- `symbol` (string, required)

Output parameters:

- `result` (object)

### `liquidations` (~326 tokens)

Liquidations bucketed by interval (1m 5m 15m 1h 4h 1d), or tick rows when interval is null.
side (the filter), position_side and long/short_liq_usd mean the liquidated position on every exchange.
A tick row's side is the exchange's raw value: sell = long liquidated on Binance, Huobi and Hyperliquid,
buy = long liquidated on Bybit, long/short on OKX. Builder plan and up. Max 5,000 rows per call.
Rows are oldest first. Without start you get the newest `limit` rows up to a fixed end (end, or
now, sent to the API) from a window that is `limit` intervals long (24 h for tick rows); with start,
the first `limit` rows from start. The result's `window` and `note` say which range was used.
Hyperliquid liquidations start 2026-05-28 (node fill data, rebuilt 10 October 2026) and usually lag 3 to 4 hours, longer if Hyperliquid's S3 uploads stall.
A dataset that's switched off returns the error code dataset_paused.
On Hyperliquid a row sums the liquidated fills sharing a time, side and price: count is rows, not fills.

Input parameters:

- `end`
- `exchange` (string)
- `interval`
- `limit` (integer)
- `side`
- `start`
- `symbol` (string, required)

Output parameters:

- `result` (object)

## Diagnostics

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

## Score history

- 2026-10-18: 59
- 2026-10-17: 59
- 2026-10-16: 59
- 2026-10-15: 59
- 2026-10-14: 34
- 2026-10-13: 34
- 2026-10-12: 34

## Common questions

### What is the CandleFeed MCP server?

CandleFeed is an MCP server listed in the public MCP registry as ai.candlefeed/candlefeed-mcp. Rebuild Binance USD-M order book days for spread and depth; candles, funding, OI and liquidations. This page covers its PyPI package (candlefeed-mcp).

### Is the CandleFeed MCP server safe to use?

CandleFeed scores 59 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 18 October 2026. 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 CandleFeed MCP server expose?

CandleFeed exposes 10 tools: l2_coverage, l2_gaps, l2_download_day, l2_book_at, l2_spread_summary, and 5 more. Their descriptions and schemas cost roughly 1,491 tokens of context every time the server is loaded.

### Is the CandleFeed MCP server still maintained?

CandleFeed is still listed as active in the MCP registry. We last reached this channel on 18 October 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 CandleFeed MCP server under?

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

## Links

- PyPI project: https://pypi.org/project/candlefeed-mcp/
- Socket report: https://socket.dev/pypi/package/candlefeed-mcp
- Website: https://candlefeed.ai/docs/order-book
- Changelog RSS feed: https://verifymcp.io/servers/ai-candlefeed-candlefeed-mcp/candlefeed-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/ai-candlefeed-candlefeed-mcp/candlefeed-mcp.json
- HTML version of this page: https://verifymcp.io/servers/ai-candlefeed-candlefeed-mcp/candlefeed-mcp