Pay-Per-Call Data API vs Subscription: Which Fits Your Usage?
Choose between a pay-per-call data API and a subscription by how your usage runs. Billing shapes, a worked per-call bill, and when a plan still wins.

Get company data with no subscription and see the price before you spend. Paste this into your agent:
Set up https://monid.ai/SKILL.md, and then use Monid to find the endpoints that enrich a company from its website, show me each one's price and billing unit before running anything, then enrich stripe.com with the one that returns its headquarters.
One prompt finds the endpoint, shows its price and runs it. 1 min setup · 6-20 hrs saved.
You need data from an API, you do not want a monthly plan for it, and you want to know what a call costs before you make it. That is the real question behind pay per call versus subscription: how your usage behaves, and whether the bill follows it. This guide covers the billing shapes, a real per-call bill from three endpoints we inspected and ran on 2026-10-16, the traps in metered pricing, and the cases where a plan wins. It runs through Monid, the OpenRouter for agent tools: one key, one balance, every tool priced per call.
Is there a data API you can pay per call instead of a subscription?
Yes: Monid sells data endpoints one call at a time, and it is the easiest way to start. You install it once, add a key, and your agent runs any endpoint on demand with no plan attached.
- Pay per call. Every endpoint shows its price and billing unit before you run it. One of the endpoints in this guide is billed at $0.00015 per call (Verified 2026-10-16 on the data tools page). Buy calls flexibly: there is no monthly plan and no minimum top-up. Discovery and inspection are free; only a run bills.
- One install, every tool. The same key and the same balance reach every endpoint, so a second data source is another call, not another vendor account.
- Lives inside your agent. Install the CLI, add your key, paste the prompt above, and the tools are there whenever your agent needs them, with nothing to keep running. Your agent reads
SKILL.mdand learns the rest, so setup is about a minute.
The sentence people use for this is "pay as you go", and it covers three different bills, which the next sections separate.
How do you buy data per call from more than one source?
Put one account in front of the sources and let each call bill on its own. The alternative is a contract per source, and the table shows how the options differ.
| Route | How it bills | What a quiet month costs | Fits |
|---|---|---|---|
| Subscription per vendor | A fixed fee per period, usage up to a cap | The full fee | Steady, high volume against one source |
| Credit pack | You buy a balance up front, usage draws it down | Nothing new, but the balance is already spent | A known volume you can pre-pay |
| Platform plan with usage on top | A base fee plus metered usage | The base fee | Teams that live on one platform |
| Monid, pay per call | A balance that moves only when a call runs, buy flexibly | Nothing | Uneven usage, many sources, agents |
The row that decides most cases is the last column: what you pay in a month when you run nothing. If you are weighing a waterfall of enrichment vendors, Clay and its alternatives shows how that bill is built. Monid bills that month at zero, because only a run bills.
For agents
Grab an API key at app.monid.ai, then paste this to your agent and hand it the key:
set up https://monid.ai/SKILL.md
It learns the whole discover, inspect, run workflow itself. More in the agent quickstart.
For humans
npm install -g @monid-ai/cli
monid keys add -k <your-api-key> -l main
The three verbs, and why two are free
monid discover -q "company enrichment" # search the catalogue, free
monid inspect -p pdl -e /v5/company/enrich # schema and price, free
monid run -p pdl -e /v5/company/enrich -i '{"website":"stripe.com"}' # the only paid step
Free discovery is what makes metered billing usable. If finding out what an endpoint costs also cost money, you could not evaluate before you commit, and the advantage of metering would be gone. On 2026-10-16 our discover search for "company enrichment" returned endpoints from Akta and Clay, each with its price and billing unit in the result.
Can you get pay as you go by signing up with each vendor yourself?
Yes, and it works for one or two sources. For each vendor you create an account, generate a key, add a payment method, read the pricing page and choose a plan or a credit pack, and then you repeat it for the next source. The cost is hours of setup per vendor and a separate bill, balance and key to track for each.
Two things make this slower than it looks. First, pay as you go means different things at different vendors, so each pricing page has to be read on its own terms: a per-call price, a per-result price and a per-section price are not interchangeable, and the Akta endpoint below uses the third. Second, the vendors that offer a metered option often sit behind a plan or a minimum purchase, so you check what a quiet month costs before you sign up.
A single account over many providers removes those steps. Monid keeps the vendor relationships and the billing, so your side is one key and one balance. The agent still chooses the endpoint; the difference is that choosing one never requires a new signup. If you later pick a source that needs a contract with seats and a service guarantee, you buy that directly and keep Monid for the rest.
Whether that tool layer beats a direct integration for an agent is its own question, covered in MCP versus a direct API for AI agents.
What does a pay-per-call bill look like?
It looks like one line per call, with the unit named. On 2026-10-16 we inspected three endpoints and ran two of them with small inputs.
A per-section call: Akta company enrichment
We asked for one section, location, of https://stripe.com from the Akta endpoint /v1/company/enrichment. The call returned HTTP 200 in about 1.3 seconds. The response carried the company uuid (0000031) and a location object: the headquarters (South San Francisco, USA, region North America), a market_served flag (is_global: true) and four offices. Each office came with a city, a country, a description, a source list of URLs and a type (headquarters, regional_hq or general_office).
The run billed one unit, 25,000 micro-dollars, which is $0.025 (Verified 2026-10-16). The endpoint is billed per section, so asking for a different section changes the price. Per monid inspect, location is the $0.025 tier and technology is $0.10, so the number of sections you request is the number you pay for. Tiered pricing like this also appears in funding data APIs and their alternatives.
A per-call endpoint: Litescrape Tripadvisor search
We searched hotels in Kyoto with place_type set to accommodation and num set to 2 on the Litescrape endpoint /tripadvisor/search. It returned total_results of 2,000 and two rows: Four Seasons Hotel Kyoto (rating 4.5, 1,065 reviews) and Hotel Granvia Kyoto (rating 4.4, 3,840 reviews), plus a ready-made pagination.next for the following page. It billed one unit, 150 micro-dollars, which is $0.00015 (Verified 2026-10-16). More detail on that endpoint is in the TripAdvisor guide.
A per-call endpoint we only inspected
The People Data Labs endpoint /v5/company/enrich is billed per call at $0.10 (Verified 2026-10-16). That is above the $0.05 ceiling we set for any single test run, so we inspected it and did not run it. Its input is a POST body with fields such as name, website, profile, ticker and min_likelihood, and its median run time per monid inspect is about 1.6 seconds.
The three lines have different units, and the difference is the lesson: one is billed per section, one per call, one per call at a higher price. A bill is only predictable once you know the unit.
What goes wrong when you meter data calls?
The unit multiplies, and the multiplication is where surprises come from. A subscription hides it inside a fixed fee; a meter shows it.
Why does the unit matter more than the price?
Because the same price buys different amounts of data. At $0.025 per section (Verified 2026-10-16), requesting one section costs $0.025 and requesting two of the $0.025 tier costs $0.05. The schema marks sections as required and tells you to list every section for the full profile, so a script that asks for all sixteen pays every tier price on every company. Read the billing line in monid inspect before you loop.
Why does an unset limit cost money?
Because per-call and per-result endpoints bill differently when a response returns many rows. The Tripadvisor search accepts num from 1 to 30 and defaults to 30; our call set it to 2 and billed one unit either way. On a per-call endpoint the page size is a page-count decision. On a per-result endpoint it is a spending control, so set it before the first run.
Why should you trust the charge and not the listing?
Because the charge is the measurement. Both runs above reported a cost that matched the listed price: 25,000 micro-dollars against $0.025, and 150 micro-dollars against $0.00015 (Verified 2026-10-16). Make a habit of checking the reported cost on your first run of any new endpoint, and compare it with the inspect price before you scale up.
What about repeated calls?
They add up linearly, and that is the point. A hundred calls at $0.00015 is $0.015, and a hundred sections at $0.025 is $2.50 (Verified 2026-10-16). The bill follows the work, which is the property a plan with a fixed fee does not have.
Which endpoint is billed in which shape?
The table lists the three endpoints from this guide with their billing unit and price. Prices were read with monid inspect on 2026-10-16.
| Endpoint | What it does | Billing | Price (Verified 2026-10-16) |
|---|---|---|---|
litescrape /tripadvisor/search | Searches Tripadvisor by text, place type and geography | Per call | $0.00015 |
akta /v1/company/enrichment | Returns the sections you pick for a company, by domain or uuid | Per section | $0.025 for the cheapest tier |
pdl /v5/company/enrich | Matches a company by name, website, LinkedIn or ticker | Per call | $0.10 |
Start with the row whose unit you understand and run one small call before you batch. The same Akta and People Data Labs pair, priced across a real list, is in what firmographics cost per record. The monid inspect output also lists the median and 95th-percentile run time, which tells you how long a loop over your list will take.
When is a subscription a better fit?
A subscription wins when your usage is steady and large, when the people using the data do not write code, or when you need a contract. We sell by the call, so check these cases before you choose.
Steady, high volume against one source. Past a certain daily load, a fixed plan spread over your records can cost less per record than any metered price. If you pull millions of records a month from one vendor, price the contract against the meter; the contract may win.
You need a UI for people who do not write code. Monid is an API, a CLI and an agent skill. If the person doing the work lives in a dashboard and a CSV, give them the vendor's own product.
You need contractual guarantees. An uptime commitment, a named account manager or signed data-processing terms come from a direct agreement with the vendor. A marketplace is built for breadth, which is a different product. We resell the endpoints in this guide, so an agreement with Monid covers the billing and key, and the data terms stay with each provider.
If you are still weighing a dataset against live calls, dataset or API: what to buy from a B2B data provider covers that fork.
Conclusion
The direct answer: pay per call fits uneven usage, many sources and agents, and a subscription fits steady volume against one source. With Monid you pay per call, buy calls as you need them with no minimum top-up, and use one key for every tool, from inside your agent. The three endpoints here bill at $0.00015 per call, $0.025 per section and $0.10 per call (Verified 2026-10-16).
Two things matter more than the model. Know the billing unit before you loop, because per call, per result and per section are different bills. And read the reported cost on your first real run, because the charge is what you pay.
How we estimated the time saved: signing up with a data vendor, creating a key, adding billing and reading the pricing page takes roughly 2 to 4 hours each, so three to five sources is about 6 to 20 hours. The 1 minute is the install, key and first call once you already have a Monid key.
Free next step: run monid discover -q "company enrichment", then monid inspect on one result to read its unit and price. Make one small call and compare the reported cost with the inspected price. Start at monid.ai.
FAQ
Is there a data API with no monthly fee?
Yes: on Monid you pay per call with no monthly plan and no minimum top-up, and discovery and inspection are free. A month in which you run nothing costs nothing, because only a run bills.
How much does one API call cost?
It depends on the endpoint and its billing unit. Through Monid the endpoints in this guide bill at $0.00015 per call, $0.025 per section and $0.10 per call (Verified 2026-10-16). monid inspect shows the price and unit for any endpoint before you run it.
What is the difference between pay per call and pay per result?
Pay per call bills once per request, however many rows come back. Pay per result bills per row or unit returned, so the page size and the number of queries set the total. The Akta endpoint in this guide is billed per section, a third shape, so read the unit first.
Is a credit pack the same as pay as you go?
Not always. A credit pack is a balance you buy up front and draw down, so you commit money before you know your volume. Pay per call on Monid is also a balance, and you buy it flexibly, with no minimum top-up, and a call draws only what that call costs.
Can I see what an API call costs before I run it?
Yes. monid inspect -p akta -e /v1/company/enrichment returns the input schema, the price and the billing unit for free. Run one small call afterwards and compare the reported cost with the inspected price.
Published August 2026.


