Akool lists API access from Pro Max at $59; Sume keys on every plan

Akool's pricing page lists API access under Pro Max at $59 a month. Sume's docs give every workspace API keys, with rate limits that start on the Free plan.

5 min readSume
All posts

Akool's pricing page lists API access as included on Pro Max, which costs $59 a month, and does not list it on Starter at $12 or Pro at $30. Sume's documentation treats API keys as a workspace feature, and its rate-limit and concurrency tables begin with a Free plan row, so an API call does not wait on a higher tier.

Akool details are from its pricing page, read 2026-10-03. Sume details are from API keys, Generation admission and Format errors and spend. Akool's page shows per-seat prices with an annual-billing equivalent, so confirm the monthly figure at checkout. A page that does not list a feature may still offer it somewhere else, so check Akool's sales terms before ruling it out.

What does each page say about the API?

On Akool's page the feature list under Pro Max reads API access included. Business adds one fine-tuned studio avatar on top, and Enterprise is custom with credits that do not expire. Starter and Pro list resolution, video length and concurrency, but no API line.

Sume's docs describe API keys as workspace-scoped: the dashboard reveals the full secret only when a key is created, and you store it in a secret manager. The plan decides limits, not whether a key exists.

API access and limits as published (read 2026-10-03)
Plan levelAkoolSume
EntryStarter $12/mo: no API line listedFree: 1 concurrent job, 120 writes and 4800 reads per minute
MiddlePro $30/mo: no API line listedPro: 4 concurrent jobs, 300 writes and 12000 reads per minute
UpperPro Max $59/mo: API access includedStartup: 8 concurrent jobs, 600 writes and 24000 reads per minute
TopBusiness $249/mo; Enterprise customScale: 20 concurrent jobs, 1200 writes and 48000 reads per minute

What a plan changes on Sume

Rate limits are per key and span all of /v1. Reads and writes have separate budgets, and the read budget is forty times the write budget, so a polling loop cannot starve your own creates. Each response carries ratelimit-limit, ratelimit-remaining and ratelimit-reset, and a 429 adds retry-after and names the budget in error.details.scope.

Request rate is not generation capacity. Raising how fast you call does not raise how many jobs run at once, which the plan's concurrency limit governs. Plan on both numbers when you size a batch.

  • Pace on the response headers rather than counting requests yourself.
  • Use Idempotency-Key on creates so a retry after 429 does not start a second job.
  • Keep the key on your server and never in a browser bundle.

When this matters

It matters when a prototype turns into an integration. If you begin on a low tier because the app is enough, a missing API line is a plan upgrade before you can automate anything. If you begin on Sume's Free plan, the first automation is a key and a curl away, and the limits you will hit first are concurrency and balance, both of which return documented errors rather than a missing feature.

Sources

Related posts

More in Comparisons

All Comparisons posts

Written by Sume