X charges for every API call now. On 2026-02-06, X retired its subscription tiers and made pay-per-use the default for anyone signing up as a new developer — no free tier, no trial reads, a credit card before your first request. The numbers below are what that actually costs, and why they still leave real X monitoring out of reach for a solo developer.
What pay-per-use actually bills you for
| Action | Price | Notes |
|---|---|---|
| Read a tweet, list, or Space | $0.005 per item | $5 per 1,000 reads |
| Read a user profile | $0.010 per item | |
| Read a like, mute, or block | $0.001 per item | |
| Post without a URL | $0.015 per post | |
| Post with a URL | $0.200 per post | 13x the cost of a link-free post |
| Follow or like | $0.015 per action |
Source: Third-party developer blogs (xpoz.ai, twitterapi.io), 2026. X's own pricing page returned a 402 error when checked directly.
Two structural limits sit on top of these per-call prices, no matter how much you're willing to pay under this tier: a hard cap of 2 million reads a month, and search that only reaches back 7 days. There is no streaming access on pay-per-use either.
The 2-million-read cap sounds generous until you notice it is a ceiling, not a knob — pay-per-use has no option to buy past it, at any price, this month or next. The 7-day search window is the sharper limit in practice: it means pay-per-use cannot answer "who has said this in the last month," only "who has said this in the last week," no matter how the query is written or how much is spent per call.
The legacy tiers nobody new can buy
Anyone who had a subscription before February 2026 is still, for now, on the old ladder — Basic, Pro, and Enterprise — though even that is being phased out. X's own developer documentation reflects the end state already: its Products navigation lists two things.
Source: docs.x.com — About the X API — captured 2026-08-20
This is the one thing in this article we could confirm against X's own site: the product structure is official. Two tiers, pay-per-use and Enterprise. What X's documentation does not publish anywhere we could find is the actual per-call rate, which is why every dollar figure in this article is attributed to third-party developer blogs rather than to X.
Source: Third-party developer blogs (xpoz.ai), 2026.
Basic is reported at $200 a month, up from an earlier $100. Pro is reported at $5,000 a month and includes full-archive search and streaming — but is closed to new signups. Enterprise runs $42,000 to $50,000-plus a month, and it is described as the only official path to full historical search back to 2006. As of 2026-06-01, X has begun moving even existing Basic and Pro subscribers onto pay-per-use, which means the subscription ladder is being retired entirely, not just closed to new customers.
One caveat that applies to everything in this section and the last, now that the screenshot has narrowed it: the product tiers are confirmed on X's own documentation, but the prices are not. Every dollar figure here — the per-read rates, the $200 Basic, the $5,000 Pro, the $42,000-to-$50,000 Enterprise band — comes from third-party developer blogs that independently report the same numbers. X's pricing page at developer.x.com returned a 402 Payment Required error every time we tried to open it, and the public documentation we could read describes what the tiers are without saying what they cost.
From free, to metered, to pay-per-use
- 2022 (pre-2023)
- A free API tier exists, used by hobbyist tools and research projects at effectively no cost.
- 2023-02
- Following the Musk acquisition, the free tier is cut down to roughly 1,500 reads a month, and a paid Basic tier launches at $100 a month.
- 2024 (through 2025)
- Basic pricing rises to $200 a month. Third-party workaround readers relying on guest access are shut off as X disables guest account creation.
- 2026-02-06
- X replaces tiered subscriptions with pay-per-use as the default for new developers — no free tier, no Basic/Pro signup option for new accounts.
- 2026-06-01
- Existing Basic and Pro subscribers begin being migrated onto pay-per-use as well.
Source: tweetapi.com, twitterapi.io, xpoz.ai — third-party developer blogs, 2026.
Why the sticker price understates the real barrier
Run the arithmetic on light use and pay-per-use looks almost reasonable: checking X once a day for mentions of your product, reading a few hundred posts each time, adds up to well under a few thousand reads a month — a few cents to a few dollars under the per-1,000-read price. Reading is not where this gets expensive.
The real ceiling shows up the moment you need anything a serious monitoring tool needs. Search only reaches 7 days back on pay-per-use, at any spend level — there is no pay-more-for-more-history option short of the closed legacy Pro tier or the $42,000-plus Enterprise tier. Writing costs scale fast too: a single link-bearing post costs 13 times what a plain one does, and following or replying to build up a real presence adds $0.015 per action on top of every read. And going outside the official API entirely, into the scraper market that grew up around X's price increases, means accepting a clear terms-of-service violation rather than a sanctioned alternative — X has already shown, by cutting off guest-account access to reader front-ends in 2024, that it will act to close informal workarounds.
None of this is a hidden fee. It's a set of walls placed exactly where a bootstrapped developer would try to build something real: a no-trial entry point, a 7-day search horizon with no paid upgrade path below five figures a month, and a terms-of-service line that scraping sits on the wrong side of.
What official access actually looks like
A handful of X-adjacent growth tools describe how they get their data. Tweet Hunter calls itself an "official X API partner," advertising "no API risk, no shadowban concerns" as a direct selling point. Typefully is listed by X itself as a documented success story on its developer site. Circleboom describes itself as an "X (Twitter) official enterprise developer" on its own help pages. Most of the smaller tools in this space don't disclose how they obtain X data at all.
That pattern — a few companies with the scale to justify an enterprise-level or officially-brokered relationship, and a long tail of tools that don't say — is itself a sign of where the price floor sits. Official, sanctioned access to X's data appears to require either enterprise-level spending or a negotiated partnership; neither is within reach of someone testing an idea on nights and weekends.
The alternative path many smaller tools take instead is routing through third-party data providers — scraper-based services that resell X data at a fraction of official pricing, running on credit systems rather than the per-call rates above. This is materially cheaper, which is exactly why it exists, and it is also a clear violation of X's terms of service rather than a gray area: X's ToS prohibits scraping outright, and X has already demonstrated it will close off informal access routes, disabling the guest-account creation that once let Nitter-style reader front-ends work around API pricing entirely. A tool built on a scraper provider is a tool built on borrowed time. That price floor is also why we sell a run instead of standing access: a WarmList run is $9, once, and nothing of ours keeps running between runs to be metered.
The actual answer to "why don't you just build it yourself"
Put the pieces together and the barrier isn't the headline per-read price — it's that there is no cheap way to try, no paid tier between $5,000 and $42,000 that includes real search history, and no legitimate discount route around the official price ladder. A solo developer who wants to watch X for people describing their exact problem is choosing between a metered tier that can't search back more than a week, a legacy tier that costs as much as a used car every month, and a terms-of-service violation.
We built WarmList because we ran into this same wall (disclosure: we build one of these) — it searches X and Reddit for people already describing the problem a product solves, without asking the person running it to sort out X's API pricing first. If you'd rather do the searching yourself by hand, Nobody's Coming. Go Reply to Them Instead. walks through exactly how, and what the reply data actually says covers how X's ranking treats a reply once you send it. For what the paid monitoring tools charge instead of building it yourself, see the X monitoring comparison.
What we charge, given all of the above
This article has priced everybody else's access to X, so it should price ours. A WarmList run is $9, paid once — no tier, no monthly minimum, no annual commitment. You paste a product URL, and the run reads public timelines, scores 500+ posts, and returns every candidate with the post, the score, and the reason it scored, roughly ten minutes after payment. That price is only possible because of what it does not include: one run, not standing access. There is no live stream to keep alive between runs and no archive to keep warm — which is precisely the thing the $5,000 and $42,000 tiers above are charging for.
