Skip to main content
Glama
CrowdCent MCP Server

CrowdCent MCP Server

Official
by crowdcent

crowdcent-mcp has moved

The CrowdCent MCP server now ships inside the crowdcent-challenge package as an optional extra. One config line replaces this repository, no clone or absolute paths needed:

{
  "mcpServers": {
    "crowdcent": {
      "command": "uvx",
      "args": ["--from", "crowdcent-challenge[mcp]", "crowdcent-mcp"],
      "env": {"CROWDCENT_API_KEY": "your_api_key_here"}
    }
  }
}

Prefer nothing installed at all? Connect the hosted server at https://mcp.crowdcent.com/mcp with your API key as a bearer token.

Migrating from this repository

If your mcp.json points at a local clone (uv run --directory /path/to/crowdcent-mcp server.py), replace that server entry with the uvx block above. That is the whole migration.

One tool changed: switch_challenge is gone. Challenge-scoped tools now take an optional challenge_slug argument instead (defaulting to hyperliquid-ranking), and new simulation tools let your assistant backtest portfolio constructions on the meta-model.

Docs: docs.crowdcent.com/ai-agents-mcp

The original source remains available in this repository's git history.

Available Tools

13 tools
download_inference_dataA
Download inference data for a specific period.

Args:
    release_date: The release date in 'YYYY-MM-DD' format or 'current' or 'latest'
    dest_path: Absolute path where to save the data, must end with .parquet
    poll: Whether to wait for the inference data to be available before downloading
    poll_interval: Seconds to wait between retries when polling
    timeout: Maximum seconds to wait before raising TimeoutError (None waits indefinitely)

Returns:
    Success message or error

Inputs: release_date (required), dest_path (required), poll, poll_interval, +1 more.

download_meta_modelB
Download the consolidated meta model for the current challenge.

Args:
    dest_path: Absolute path where to save the meta model, must end with .parquet

Returns:
    Success message or error

Inputs: dest_path (required).

download_training_datasetB
Download a specific training dataset.

Args:
    version: The version string of the training dataset (e.g., '1.0', '2.1') or 'latest'
    dest_path: Absolute path where to save the dataset, must end with .parquet

Returns:
    Success message or error

Inputs: version (required), dest_path (required).

get_challenge_infoB
Get detailed information about the current challenge.

Returns:
    Dictionary containing challenge details

Inputs: no named parameters.

get_inference_data_infoB
Get detailed information about a specific inference data period.

Args:
    release_date: The release date in 'YYYY-MM-DD' format or 'current' or 'latest'

Returns:
    Dictionary containing inference data details

Inputs: release_date (required).

get_submissionC
Get details about a specific submission.

Args:
    submission_id: The ID of the submission

Returns:
    Dictionary containing submission details

Inputs: submission_id (required).

get_training_dataset_infoB
Get detailed information about a specific training dataset.

Args:
    version: The version string of the training dataset (e.g., '1.0', '2.1') or 'latest'

Returns:
    Dictionary containing dataset details

Inputs: version (required).

list_all_challengesB
List all available challenges.

Returns:
    Dictionary containing list of challenges with their details

Inputs: no named parameters.

list_submissionsA
List recent submissions.

Args:
    period: Optional filter for submissions by period:
            - 'current': Only show submissions for the current active period
            - 'YYYY-MM-DD': Only show submissions for a specific inference period date

Returns:
    Dictionary containing list of submissions

Inputs: period.

list_training_datasetsB
List all available training datasets for the current challenge.

Returns:
    Dictionary containing list of training datasets

Inputs: no named parameters.

submit_predictions_from_dataframeC
Submit predictions from a JSON representation of a dataframe.

Args:
    df: dataframe containing predictions data
    slot: Submission slot number (1-based, default: 1)

Returns:
    Dictionary with submission details

Inputs: df (required), slot.

submit_predictions_from_fileB
Submit predictions from a Parquet file.

Args:
    file_path: Absolute path to the predictions file, must end with .parquet
    slot: Submission slot number (1-based, default: 1)

Returns:
    Dictionary with submission details

Inputs: file_path (required), slot.

switch_challengeC
Switch to a different challenge.

Args:
    challenge_slug: The slug of the challenge to switch to

Returns:
    Success message

Inputs: challenge_slug (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 13 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose targeting specific resources and actions. For example, download_inference_data vs get_inference_data_info provide complementary but non-overlapping functionality, and the two submission tools handle different input formats without ambiguity.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern with snake_case throughout. The naming convention is predictable: download_*, get_*, list_*, submit_*, and switch_* prefixes clearly indicate action types, making the tool set highly readable and systematic.

    Tool Count5/5

    With 13 tools, this server is well-scoped for its apparent domain of challenge management and data handling. Each tool serves a specific purpose in the workflow, from data retrieval to submission management, without redundancy or obvious omissions in the count.

    Completeness4/5

    The tool set provides comprehensive coverage for core challenge workflows, including data access, information retrieval, submission handling, and challenge switching. A minor gap exists in update/delete operations for submissions or challenges, but agents can work effectively with the available CRUD-like operations.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers