Why I built GetXAPI: the Twitter API math that made me angry

#api #twitter #ai #buildinpublic

Originally published on GetXAPI Substack. Cross-posted here for the dev community. Follow @GetXAPI for the next build log.

The bill that started this

I needed Twitter data at scale for a project in early 2025. I priced what it would cost on the official Twitter / X API.

$5,000 a month. Pro tier. 1M tweets included.

For a project that was making zero dollars and would never make $5K a month, that was the end of it. Not because the data was not valuable. Because the API math was broken.

I checked the alternatives. TwitterAPI.io was $0.15 per 1,000 tweets, roughly $150 for the same 1M tweets. Apify actors ran $0.40 to $1.50 per 1,000 tweets. RapidAPI wrappers were all over the place ($0.30 to $2.00 per 1,000).

Nothing fit. So I built GetXAPI.

What GetXAPI is

GetXAPI is a pay-per-call Twitter / X API with 44 endpoints. $0.001 per call. That works out to $0.05 per 1,000 tweets read. Same 1M tweets from the example above costs $50 a month instead of $5,000.

The pricing is not a promotion. There is no monthly minimum. There is no per-developer key fee. There is no 7-day approval queue. You hit signup, get $0.5 of free credit (no credit card), grab a bearer token, and you are running queries in under 60 seconds.

The 44 endpoints cover both reads (search, profiles, follower graph, timeline, bookmarks, mentions, lists, communities, trends) and writes (post tweets, post replies, like, retweet, bookmark, follow, send DMs, post articles, update profile). The full machine-readable spec is at docs.getxapi.com/openapi.json. The interactive docs are at Fern. There is a Postman collection with 44 working request examples.

The cost math at three volumes

This is the table I wish I had when I was deciding. All prices in USD per month for read-only workloads. Write endpoints add cost differently per provider; I will write a separate post on that.

At 100K tweets per month: GetXAPI $5. TwitterAPI.io $15. Apify (mid-range $0.80 per 1k) $80. Official Twitter API Basic $200 (subscription minimum, even if you read just 1 tweet). Official Twitter API Pro $5,000 (subscription minimum).

At 1M tweets per month: GetXAPI $50. TwitterAPI.io $150. Apify $800. Official Twitter API Basic not possible (50K cap). Official Twitter API Pro $5,000.

At 10M tweets per month: GetXAPI $500. TwitterAPI.io $1,500. Apify $8,000. Official Twitter API Basic not possible. Official Twitter API Pro not possible (1M cap). Official Twitter API Enterprise: starts at $42,000 per month, custom contract required.

The pricing gap is not a small optimization. At 1M tweets, GetXAPI is 100x cheaper than the Twitter Pro tier. At 10M tweets, the official API has no real-world answer; the Enterprise contract is 84x more expensive than GetXAPI.

What shipped in the first month

The 44-endpoint coverage and the $0.001 per-call pricing were the table stakes. The interesting work was the distribution surface, because a Twitter / X API only matters if developers can actually find it and wire it into the tools they already use.

The GetXAPI MCP server shipped to npm at @getxapi/mcp. Run npm install -g @getxapi/mcp and every endpoint becomes a tool inside Claude Desktop, Cursor, Continue, or any other MCP-compatible AI client. This is the surface I am most excited about. Wiring a Twitter / X API into an AI agent in two commands beats every other developer-experience moment in the API the team has shipped this year.

The GetXAPI Apify Store shipped with the first actor at create-tweet. Apify actors are the no-code path: scheduled runs, webhooks, retry handling, all managed by Apify. More actors are landing monthly as the team migrates the most-requested endpoint workflows into actor form.

The docs ship through Fern (live) for the interactive reference experience and Postman (live) for the import-and-run experience. The OpenAPI 3.1 spec at docs.getxapi.com/openapi.json is the source of truth that both surfaces consume.

What I am writing about here

This Substack is the build log. Field notes on running a production Twitter / X API. Comparison teardowns of every Twitter / X data option in the market (one per quarter, with the math). Tutorial posts on wiring GetXAPI into the tools developers already use (Claude, Cursor, LangChain, LlamaIndex, ElizaOS, MCP, raw REST). Build-log updates on what shipped each week and why. Opinion posts on the state of Twitter / X data access in 2026.

The next post (Friday) is the tutorial on wiring the GetXAPI MCP server into Claude Desktop in five minutes. The post after that is the cost-math deep dive on write endpoints. Subscribe below if you want the build log delivered.