name: BigML Rate Limits
description: Rate limits and task quotas for the BigML API
url: https://bigml.com/pricing

notes: >
  BigML enforces limits primarily via parallel task capacity and dataset size
  constraints, which vary by subscription tier. Specific HTTP-level rate limits
  per second are not publicly documented; contact BigML support for enterprise
  limits. Use api.get_tasks_status() to check current task counts and availability.

limits:

  - tier: Free
    parallel_tasks: 2
    max_dataset_size: 16 MB per task
    notes: Shared infrastructure, basic priority queue

  - tier: Prime
    parallel_tasks: Configurable (based on subscription)
    max_dataset_size: Configurable (based on subscription)
    notes: Prioritized job queues, flexible server topology

  - tier: Private Deployment
    parallel_tasks: Unlimited (constrained by server cores)
    max_dataset_size: Unlimited (constrained by server resources)
    notes: Dedicated servers; 1 server = 8 cores for Lite and Bronze tiers

status_check:
  method: api.get_tasks_status()
  description: Returns current task counts, maximums, and availability metrics
  fields:
    - tasks: current number of running tasks
    - max_tasks: maximum allowed parallel tasks
    - available: whether new tasks can be submitted

regional_endpoints:
  - region: Global
    base_url: https://bigml.io
  - region: Australia (VPC)
    base_url: https://au.bigml.io
  - region: Virtual Private Cloud
    base_url: Configured per deployment