specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: ForceDream
providerId: forcedream-ai
generated: '2026-09-19'
method: searched
source: https://www.forcedream.com/developers/rate-limits
created: '2026-09-19'
modified: '2026-09-19'
tags:
  - Rate Limiting
  - AI Agents
description: >-
  ForceDream documents token-bucket rate limits per subscription tier with four response headers and a 429 on
  exhaustion. Several other limits are published elsewhere and recorded here too: a 60 requests/minute
  per-account limit on the A2A runtime, a 5-per-IP-per-hour signup limit declared in the agent card, a
  100-calls/hour sandbox limit, and a daily earn limit that resets at midnight UTC. The MCP endpoint returned
  X-RateLimit-Limit: 120 to an anonymous client during profiling — a value that appears on no documented tier
  and is recorded as observed.
sources:
  - https://www.forcedream.com/developers/rate-limits
  - https://github.com/forcedreamai/forcedream-a2a-agent/blob/main/examples/retry-backoff.md
  - https://api.forcedream.ai/.well-known/agent-card.json
  - https://www.forcedream.com/sandbox
  - https://www.forcedream.com/developers/errors
  - live response headers from https://api.forcedream.ai/v1/mcp, 2026-09-19
algorithm: token bucket
headers:
  limit: X-RateLimit-Limit
  remaining: X-RateLimit-Remaining
  reset: X-RateLimit-Reset
  retryAfter: Retry-After
  header_semantics:
    X-RateLimit-Limit: Requests allowed per minute
    X-RateLimit-Remaining: Requests remaining in current window
    X-RateLimit-Reset: Unix timestamp when window resets
    Retry-After: Seconds to wait if rate limited (429 responses only)
responseCodes:
  throttled: 429
  dailyLimit: 429
limit_count: 9
limits:
  - name: Free tier
    scope: account
    metric: requests_per_minute
    limit: 60
    timeFrame: minute
    burst: 10
    concurrent: 2
    daily: 1000
  - name: Starter tier
    scope: account
    metric: requests_per_minute
    limit: 300
    timeFrame: minute
    burst: 50
    concurrent: 5
    daily: 10000
  - name: Pro tier
    scope: account
    metric: requests_per_minute
    limit: 1000
    timeFrame: minute
    burst: 200
    concurrent: 20
    daily: 50000
  - name: Scale tier
    scope: account
    metric: requests_per_minute
    limit: 5000
    timeFrame: minute
    burst: 1000
    concurrent: 100
    daily: 500000
  - name: Enterprise tier
    scope: account
    metric: requests_per_minute
    limit: -1
    timeFrame: minute
    note: '"Unlimited" requests; burst and concurrency "Custom"'
  - name: A2A runtime per-account limit
    scope: account
    metric: requests_per_minute
    limit: 60
    timeFrame: minute
    error_code: rate_limit_exceeded
    source: https://github.com/forcedreamai/forcedream-a2a-agent/blob/main/examples/retry-backoff.md
  - name: Self-service signup
    scope: ip
    endpoint: POST /api/signup
    metric: requests_per_hour
    limit: 5
    timeFrame: 3600 seconds
    source: agent card self-service-credentials extension
  - name: Hosted sandbox
    scope: shared-test-account
    metric: calls_per_hour
    limit: 100
    timeFrame: hour
    source: https://www.forcedream.com/sandbox
  - name: Daily earn limit
    scope: account
    metric: earn_events_per_day
    limit: null
    timeFrame: day
    note: '"Daily limit reached" -> 429; resets at midnight UTC. The numeric ceiling is not published.'
observed:
  - endpoint: POST https://api.forcedream.ai/v1/mcp (anonymous)
    headers: {x-ratelimit-limit: '120', x-ratelimit-remaining: '119', x-ratelimit-reset: '1789873576'}
    note: Limit value not on any documented tier; window inferred from reset (unix seconds) not stated.
  - endpoint: GET https://api.forcedream.ai/v1/agents/lead-score-v1 (401) and GET /v1/account/balance (401)
    headers: none
    note: Unauthenticated REST rejections carried no rate-limit headers.
backoff_guidance: '"Set backoffMs to at least 1000ms and maxBackoffMs to several seconds so retries land in the next rate-limit window" (A2A retry guide); docs example honours Retry-After then falls back to 2^attempt seconds.'