X API Goes Pay-Per-Use: What Indie Hackers Should Build Now

Starting January 21, 2026, X transitioned its entire API to usage-based pricing. No more monthly subscriptions. No more rate-limit tiers. You buy credits in the Developer Console, and they're deducted as you make requests.

The shift was telegraphed. X ran a closed beta starting October 2025, handing out $500 vouchers to selected developers. By February, the old fixed plans were removed from the pricing documentation entirely. Existing users on fixed plans can still access them through the dashboard—but it's only a matter of time before everyone migrates.

This pricing model isn't universally cheaper. It's a bet on attracting a large volume of small developers while extracting more from heavy users. Here's where the math lands:

For the past two years, X's API pricing was a wall. The $200/month minimum meant that any project built on X data needed to generate revenue from day one just to cover API costs. Side projects died. Bots went offline. Beloved third-party apps like Tweetbot and Twitterrific shut down permanently.

Pay-per-use changes the math entirely. You can now validate an idea on X data for $5–$50/month. Build an MVP, test it with real users, and only scale your API spending when your product actually generates revenue. That's the kind of economics that lets indie hackers experiment.

X itself framed this as targeting "indie builders, early-stage products, startups, and hobbyists." Whether they follow through on developer relations remains to be seen, but the pricing finally makes it possible to build without committing $2,400/year upfront.

The pay-per-use model creates a specific sweet spot: tools that need moderate read access (under 50,000 requests/month) and deliver enough value to charge $29–$99/month. Here are five ideas that fit.

Here's the business model shift most people are missing. Pay-per-use pricing doesn't just make it cheaper for developers to build on X—it creates a new product category: BYOK (Bring Your Own Key) tools sold as one-time purchases.

Think about it. Existing X tools like Typefully, Hypefury, and Tweet Hunter charge $29–$49/month because they're absorbing the API costs on your behalf. Those costs were fixed at $200/month minimum, so the SaaS model was the only way to make the math work. But now that X charges per request, a developer can ship a self-hosted or desktop app where the user plugs in their own X API key—and charge a one-time fee instead.

The user pays X directly for their usage (which might be $5–$30/month depending on volume), and pays the tool creator once. No recurring fees. No middleman markup on API costs. The tool creator has zero marginal cost per user and keeps 100% of the sale.

Before you go all-in building on X's API, the elephant in the room: trust. X killed free API access overnight in 2023. They doubled Basic pricing from $100 to $200. Third-party apps that had served millions of users for over a decade were wiped out with weeks of notice.

Developers who got burned once are right to be cautious. The question isn't "Is this pricing good?" but "Will it stay stable?" Building a business on a platform that has changed its API terms this aggressively is a calculated risk.

X's move to pay-per-use is part of a broader trend. API pricing across the industry is shifting from fixed tiers to usage-based models—OpenAI, Twilio, and Stripe all price this way. For developers, it means lower barriers to entry but higher costs at scale.

The xAI integration is the most interesting signal. By connecting API spending to AI credits, X is positioning itself not just as a social platform with an API, but as a combined social + AI development platform. If that integration deepens—imagine Grok models trained on real-time X data available to API developers—the opportunity gets significantly bigger.