Plans lists 14,861 pricing documents, and we can read 14,857 of them across 12,992 provider folders. Each one carries a method field, and that field matters more than any price in the file. 7,611 records were searched from the provider’s own page. 4,725, or 32%, were generated by us. 7,727 are marked evidence_backed: true, and not one of the generated ones is.
| Method | Records | Evidence backed |
|---|---|---|
| searched | 7,611 | 7,605 |
| generated | 4,725 | 0 |
| estimated | 1,808 | 0 |
| none (a recorded absence) | 503 | 0 |
| probed | 121 | 121 |
| no method field | 58 | 0 |
| derived, unmarked, provider-published | 31 | 1 |
The other four links on the page are malformed. Three have an apis.yml URL pasted into the path, and one has a stray space in the slug. Those are ours.
Who publishes what
Depth varies widely. SignalWire has 64 plans in one record, GoodUnited 50, Stripe 48 and ThingsBoard 41. At the other end, 4,626 records hold no plans at all, and 1,805 of those are estimate files, which carry formulas in place of tiers. 3,059 records have prices_published set to true and 1,273 have it set to false, which is a finding in its own right. The other 10,525 have no value in that field.
audience is the field an integrator should filter on first. Where it is set, 3,286 records price a business product and 1,860 price a developer API. A subscription price is not the cost of an API call, and the file says so.
What a good one looks like
The Zoho CRM record is worth having because it transcribes what a program needs: api_credits_per_day and api_concurrency per tier. Free gets 5,000 credits a day and a concurrency of 5. Enterprise gets “50000 + (users x 1000) + add-ons, capped at 5000000” and a concurrency of 20. That is capacity planning, not marketing.
Stripe’s record is method: searched, pricing_basis: ad-valorem, with a date and a source URL, and its Standard plan reads “2.9% + 30¢” per successful transaction for domestic cards. Infobip’s is more useful for its honesty. Its note says the page uses interactive per-country calculators that show only €0 placeholders in the text, so the only explicit numbers captured are Viber minimum commitments of 175 and 115 euros. A record that tells you what it could not read saves you from assuming the rest is free.
The Datadog estimates file shows the discipline we want from derived data. It opens “ESTIMATES, not quotes,” says each estimate names the figure it came from, and says the estimate retires once the provider publishes that field.
Why an integrator or agent cares
An agent choosing between two providers will read a price if it is there. If it cannot tell a transcribed price from a generated plan shape, it will reason from numbers nobody quoted. 3,932 of the generated records follow the API Commons plan shape and are useful as a map of what tiers exist, such as Webflow’s 14 plans. They are not a quote, so an agent should filter on method and evidence_backed before comparing cost.
The work left is ours. The 4,725 generated records should be searched or probed against the provider’s page, and prices_published should be filled in across the 10,525 records that lack it.
Takeaway
Treat a pricing record as a claim with a source, and check method before you trust the number. Browse them at apis.io/plans/.