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: Algorand
providerId: algorand
created: '2026-06-13'
modified: '2026-06-13'
reconciled: true
tags:
  - Blockchain
  - Proof-of-Stake
  - Rate Limiting
  - API Access
description: >-
  Rate limits for Algorand Algod and Indexer REST APIs depend on the infrastructure provider. The
  Algorand daemons themselves impose no hard rate limits when self-hosted; limits are enforced by
  managed service providers such as Nodely. The free tier (no API key) applies a 60 req/s per
  browser/frontend ceiling and global monthly quotas for backend traffic. Commercial tiers offer
  substantially higher throughput with per-key and per-IP limits.
sources:
  - https://nodely.io/
  - https://nodely.io/docs/free/start/
  - https://nodely.io/docs/free/endpoints/
responseCodes:
  throttled: 429
  quotaExceeded: 429
limits:
  - name: Free tier frontend requests per second
    scope: ip
    metric: requests_per_second
    limit: 60
    timeFrame: second
    notes: >-
      Applies to browser/frontend traffic on the free public endpoints (no API key).
      Backed by 50 ms artificial latency injection per response.
  - name: Free tier monthly request quota
    scope: global
    metric: requests_per_month
    limit: 6000000
    timeFrame: month
    notes: >-
      Aggregate monthly cap shared across all free-tier users for backend/bot traffic.
      Frontend browser traffic has a separate 150 GB data transfer cap.
  - name: Free tier monthly data transfer
    scope: global
    metric: gb_per_month
    limit: 150
    timeFrame: month
    notes: Frontend bandwidth cap on the free tier; applies to browser-origin requests.
  - name: Unlimited tier requests per second per key
    scope: key
    metric: requests_per_second
    limit: 6000
    timeFrame: second
    notes: Per API key or whitelisted domain on the Unlimited commercial plan.
  - name: Unlimited tier requests per second per IP
    scope: ip
    metric: requests_per_second
    limit: 500
    timeFrame: second
    notes: Per IP address ceiling on the Unlimited commercial plan.
  - name: Business SLA tier requests per second per IP
    scope: ip
    metric: requests_per_second
    limit: 500
    timeFrame: second
    notes: Per IP ceiling on the Business SLA annual plan.
  - name: Business SLA tier daily request quota
    scope: account
    metric: requests_per_day
    limit: 200000
    timeFrame: day
    notes: Daily request cap on the Business SLA plan; resets at midnight UTC.
policies:
  - name: Retry on 429
    description: >-
      On receiving a 429 Too Many Requests response, back off exponentially with jitter before
      retrying. Nodely does not document a Retry-After header, so implement client-side
      exponential backoff starting at 1 second.
  - name: Pagination for Indexer queries
    description: >-
      Use cursor-based pagination (next-token) when querying large transaction or account result
      sets via the Indexer API. Paginating in moderate page sizes avoids spiking request rates
      and keeps individual responses fast.
  - name: Use testnet for development
    description: >-
      Run development and integration tests against the free testnet endpoints
      (testnet-api.4160.nodely.dev, testnet-idx.4160.nodely.dev) to avoid consuming mainnet
      quota and to ensure idempotent testing with faucet funds.
  - name: LocalNet for unit testing
    description: >-
      Use AlgoKit LocalNet (a local sandboxed Algorand network) for unit and simulation tests
      that do not require live network data, eliminating all network rate-limit concerns.
  - name: Self-host for unrestricted throughput
    description: >-
      Operators requiring unlimited throughput can run their own Algod and Indexer daemons.
      The software is open-source (go-algorand, algorand/indexer) and imposes no built-in
      per-request rate limits.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com