Flat Files MCP Tool Reference

The Flat Files MCP server currently exposes 2 tools.

Bucket discovery

storage_list_buckets

List all storage buckets available to the authenticated API key.

Arguments

This tool does not require any arguments.

Returns

The response includes:

• buckets, an array of bucket objects
• owner, metadata about the bucket owner

Each bucket item can include fields such as:

• name
• creationDate
• bucketRegion

Use when

• You want to confirm which buckets are visible to the current API key.
• You need the bucket name before calling storage_list_objects.
• You want to validate MCP connectivity with the simplest possible request.

Object and prefix discovery

storage_list_objects

List objects in a bucket with an optional prefix filter. The response follows an S3-style layout and can represent either folder-like prefixes or concrete files.

Required arguments

Optional arguments

Returns

Depending on the requested level, the response can include:

• isTruncated
• contents
• prefix
• maxKeys
• commonPrefixes
• delimiter
• keyCount
• contentLength
• httpStatusCode

When contents is present, each item can include:

• key
• lastModified
• size

When commonPrefixes is present, each item includes:

• prefix

Use when

• You want to inspect the bucket hierarchy without downloading files.
• You need to discover dataset families such as T-TRADES/ or T-LIMITBOOK_FULL/.
• You want to enumerate date partitions under a known dataset prefix.
• You need file metadata like object key, last modified timestamp, and size before downloading through S3.

Practical examples

Example: list all accessible buckets

{
  "tool": "storage_list_buckets",
  "arguments": {}
}

Example: list top-level dataset prefixes in coinapi

{
  "tool": "storage_list_objects",
  "arguments": {
    "bucket": "coinapi"
  }
}

This currently returns commonPrefixes such as:

• T-TRADES/
• T-QUOTES/
• T-LIMITBOOK_FULL/
• T-LIMITBOOK_SNAPSHOT_50/
• T-LIMITBOOK_DEPTH_BANDS/
• T-OHLCV_ACTIVE_CONSOLIDATED/

Example: list yesterday's daily partition in coinapi-daily-tail

The coinapi-daily-tail bucket holds the previous day's daily partition only (24-hour retention). Use D-YYYYMMDD where the date is yesterday (UTC). OHLCV is not available in this bucket.

{
  "tool": "storage_list_objects",
  "arguments": {
    "bucket": "coinapi-daily-tail",
    "prefix": "T-TRADES/"
  }
}

Live-tested results currently include prefixes such as:

• T-TRADES/D-20260609/

Example: list hourly partitions for trades in coinapi

Starting from 2026-06-09, new Trades, Quotes, and Full Limit Order Book data in coinapi uses hourly partitions (D-YYYYMMDDHH). Legacy daily partitions published before that date remain in coinapi under D-YYYYMMDD.

{
  "tool": "storage_list_objects",
  "arguments": {
    "bucket": "coinapi",
    "prefix": "T-TRADES/"
  }
}

Live-tested results currently include prefixes such as:

• T-TRADES/D-2026040712/
• T-TRADES/D-2026040711/
• T-TRADES/D-2026040710/

Example: continue drilling into a prefix

{
  "tool": "storage_list_objects",
  "arguments": {
    "bucket": "coinapi",
    "prefix": "T-LIMITBOOK_FULL/"
  }
}

Use the same tool repeatedly with narrower prefixes until you reach concrete object keys in contents.

Usage guidance

• Start with storage_list_buckets, then choose coinapi for hourly historical access and OHLCV, or coinapi-daily-tail for the previous day's daily files only.
• Use broader prefixes first and narrow them gradually instead of guessing full file keys.
• Expect folder-like results to appear in commonPrefixes and actual files to appear in contents.
• Keep the trailing slash when browsing hierarchical prefixes, for example T-TRADES/.
• Use MCP for discovery and the standard S3 API for file downloads.