# OpenTweet Aiolum story: https://aiolum.com/story/opentweet/ · added 2026-10-27 · updated 2026-10-27 **In short.** OpenTweet (Posting layer for X (Twitter): an MCP server, REST API, CLI and scheduler that lets a person or an AI agent post to X without an X developer account, Serbia): US$4K MRR verified by Paddle, October 2026. Evidence grade A: disclosed by a company or a filing, or verified by a payment processor. Team: 1 person (TrustMRR), the founder known as Branko; no other headcount published. Launched December 2025; still running, revenue growing. Posting layer for X (Twitter): an MCP server, REST API, CLI and scheduler that lets a person or an AI agent post to X without an X developer account · Serbia · AI-native · solo founder · still running, revenue growing · group: Marketing and social content - Outcome: US$4K MRR verified by Paddle, October 2026 (evidence grade of the headline number: A; headline in US dollars: 48300) - Launch: December 2025; months from launch to the outcome: 10 - Team: 1 person (TrustMRR), the founder known as Branko; no other headcount published - Outside funding: Bootstrapped (TrustMRR) - Main factor: Timing - Meters, 1 to 4: evidence strength 3, copyable pattern 2, capital needed 1 - Tags: AI-native, Solo founder, Bootstrapped, Running, B2B, Marketing **The product.** A hosted service that connects a person's own X account with one click, then lets them, their code or an AI agent post to X through a Model Context Protocol (MCP) server, a REST API and a command-line tool, with no X developer account. The same account has a calendar, an AI writer trained on the user's voice, an evergreen queue that re-posts the user's best tweets, connectors that turn RSS, GitHub or Stripe events into drafts, and analytics. It also cross-posts the same draft to Bluesky and LinkedIn. Everything goes through the official X API: it does not scrape, drive a logged-in browser or use the user's password. Plans are US$19.99, US$29 and US$49 a month. **Why it worked.** In February 2026 X turned its API into a pay-per-use meter, charging about US$0.015 a post and US$0.20 for a post with a link, and asking every developer to run their own X developer account. OpenTweet sits on top of that: the user connects their account once, and then posts from an agent, a script or a calendar for a flat monthly price. TrustMRR, reading the company's Paddle account, shows US$4,025 of monthly recurring revenue from 231 subscriptions and US$26,888 of revenue in total on 27 October 2026. Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; calc = computed on this page from graded facts. ## Summary OpenTweet is a posting layer for X (formerly Twitter): a service that holds the connection to a user's X account and offers a single way to publish to it, so that a program never talks to X directly. It is built around the Model Context Protocol (MCP), an open standard from Anthropic that lets an AI assistant call outside tools; OpenTweet's MCP server gives an agent 43 tools, such as create_tweet and schedule_tweet. The same account also has a REST API (a web address a program sends requests to) and a command-line tool, plus a human calendar, an AI writer and analytics. A user connects their own X account with one click, so OpenTweet needs no X developer account, the paid, metered access that X requires of anyone who calls its API directly. The founder is known only as Branko (the TrustMRR founder link is "brankopetric00"); the TrustMRR profile gives the team size as 1 person, the funding as bootstrapped, and the country as Serbia. TrustMRR, a site that reads revenue directly from a company's payment account through an API key, shows for OpenTweet's Paddle account (Paddle is a payment service that also handles sales tax as the seller of record) US$26,888 of revenue in total, all of it in the last twelve months, and monthly recurring revenue (MRR, the monthly value of all active subscriptions) of US$4,025 from 231 active subscriptions on 27 October 2026. Monthly revenue climbed from almost nothing in December 2025 to about US$3,200 to US$4,400 a month over the summer of 2026. The revenue figures are grade A: TrustMRR reads them from the company's Paddle account. The founder, the team size and the funding come from the founder-written part of the TrustMRR profile, grade C. The page also shows earlier companies that built tools on top of social-media platforms and their APIs, and died when the platform changed its rules, which is the main risk to a product that lives on X's API. ## Pricing OpenTweet prices on 27 October 2026, from the pricing page and the home page (grade C). The annual per-month figure is calculated on this page | Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit | |---|---|---|---| | Pro | 16.59 | 19.99 | 1 X account, 20 posts a day, 1 link post a day, 10 AI generations a day, 3 API keys | | Advanced | | 29 | 3 X accounts, 100 posts a day, 10 link posts a day, 50 AI generations a day, 10 API keys, analytics | | Agency | | 49 | 10 X accounts, 300 posts a day, 20 link posts a day, 150 AI generations a day, 25 API keys, analytics | All three plans include the REST API, the CLI and the hosted MCP server with 43 tools, cross-posting to Bluesky and LinkedIn, an evergreen queue and AI writing. The difference is volume: posts a day, link posts a day, AI generations a day, the number of X accounts and API keys. Annual billing saves a stated 17% (the home page shows US$15.83 a month for Pro billed yearly). Every plan has a 7-day free trial that needs a card, and a 7-day money-back guarantee. There is no permanent free plan. ## How much revenue does OpenTweet make? The numbers, graded US$4,025 of Paddle-verified monthly recurring revenue from 231 subscriptions and US$26,888 of revenue in total on 27 October 2026, with monthly revenue up from almost nothing in December 2025 | Fact | Value | Grade | Note | Sources | |---|---|---|---|---| | Monthly recurring revenue (MRR), 27 October 2026 | US$4,025 from 231 active subscriptions | A | TrustMRR, read from the company's Paddle account through an API key, synced 27 October 2026 at 03:44 UTC. 30-day MRR growth: +7.3%, so it was about US$3,751 thirty days before (calculated: 4,025 / 1.073). | [1] | | Run rate from the MRR | about US$48,300 a year | calc | Calculated: 4,025 times 12. This is the outcome amount stored for this page. The last 30 days of revenue, US$4,198, give a similar run rate of about US$50,376. | [1] | | Revenue per subscription per month | about US$17.42 | calc | Calculated: 4,025 / 231. Close to the US$19.99 Pro plan, so most subscriptions are the cheapest plan, billed monthly or at the lower annual rate. | [1, 3] | | Verified revenue, all time, to 27 October 2026 | US$26,888 | A | TrustMRR, all of it in the last 12 months. Last 6 months: US$21,854; last 3 months: US$11,771; last 30 days: US$4,198, marked +1.7% against the 30 days before. | [1] | | Revenue by month | Dec 2025 US$10; Jan 2026 US$27; Feb US$207; Mar US$2,078; Apr US$2,803; May US$3,700; Jun US$3,238; Jul US$3,359; Aug US$3,628; Sep US$4,384; Oct US$3,455 to 27 October | A | TrustMRR monthly table, synced 27 October 2026. A fast rise from March to May, then roughly flat around US$3,200 to US$4,400; September was the best month so far. | [1] | | Where the revenue figures come from | Paddle, through an API key; merchant of record: no | A | TrustMRR's verification section. Google Analytics is also connected, so the traffic figure is measured: 45,597 visitors in the last 30 days, a revenue per visitor of US$0.09. | [1] | | Founder, team and funding | 1 person, known as Branko; bootstrapped | C | TrustMRR founder-written fields: "Team size: 1 person", "Funding: Bootstrapped", founder link "brankopetric00", X followers 20,560. No full name and no employee is named in the sources read. | [1] | | Founded date | 1 December 2025 | C | TrustMRR founder-written field. The first verified revenue is from December 2025. | [1] | | Estimated users | 14,000; the site says "14,000+ users" and "80,000+ posts by agents" | C | TrustMRR founder-written estimate and the home page. These are not payment-verified; the verified figure is 231 paying subscriptions. | [1, 3] | | Why it exists: X's API became a meter | about US$0.015 a post, US$0.20 for a post with a link, since February 2026 | C | Home-page FAQ: "The official X API and XMCP both require an X developer account and pay-per-use API pricing since February 2026". OpenTweet charges a flat monthly fee instead and includes a few link posts a day. | [3] | | No auto-replies to other people | blocked by X since 23 February 2026; OpenTweet does not do it | C | FAQ: "Can my agent reply to other people's posts on X? No... On 23 February 2026 X restricted programmatic replies... only Enterprise access is exempt. OpenTweet posts through the official X API, so the restriction applies to us as well." The Replies feature drafts a reply to a post the user picks and the user sends it from X's own composer. | [3, 6] | | Stated compliance with X's rules | official API only; no scraping, no browser automation, no password | C | FAQ: "Scheduled and automated posting is allowed under X's automation rules as long as it goes through the official API on an account you authorized... It does not scrape, it does not drive a logged-in browser session, and it never signs in with your password." | [3] | | DM outreach is human-approved | candidates from real engagement, drafted, then stopped until the user approves | C | Home page, DM Campaigns: "OpenTweet finds candidates from real engagement, drafts each message, and then stops. Nothing sends until you approve it, and every send goes through the official API rather than a scripted browser." | [3] | | Networks and formats | posts, threads up to 25 parts and Articles on X; the same draft cross-posts to Bluesky and LinkedIn | C | Home page. Threads is on the roadmap; Instagram and TikTok are "not planned". Analytics are X-only. | [3] | | AI writing models | seven models across Anthropic, OpenAI, Google and xAI | C | Home-page FAQ: "OpenTweet runs your prompt through any of seven models across Anthropic, OpenAI, Google and xAI... Nothing publishes until you review, edit and approve it." The pricing page names Claude, GPT and Gemini models. | [3, 4] | How to read this table: a grade A value was read by TrustMRR from the company's Paddle account through an API key. A grade C value is what the company states on its own pages or in the founder-written part of its profile. An "X" value was computed on this page from the rows above it. ## When did OpenTweet launch, and how did it grow? Timeline From a product founded in December 2025 to US$26,888 of Paddle-verified revenue ten months later. Each entry carries its grade. - **1 December 2025** (grade C [1]): Founded date on the TrustMRR profile. First verified revenue the same month: about US$10. - **February 2026** (grade C [3, 6]): X turns its API into a pay-per-use meter and restricts programmatic replies on self-serve tiers (23 February). OpenTweet's flat-price model answers the first and it drops auto-replies because of the second. - **March 2026** (grade A [1]): Verified revenue reaches US$2,078, up from US$207 in February, the month the product starts to sell. - **May 2026** (grade A [1, 3]): Verified revenue US$3,700. The paid tiers of X's own API (US$200 and US$5,000) are retired to new signups this month, by OpenTweet's FAQ. - **September 2026** (grade A [1]): Best month so far: US$4,384 of verified revenue. - **27 October 2026** (grade A [1, 3]): TrustMRR shows US$4,025 of MRR from 231 subscriptions and US$26,888 in total. The site lists 14,000+ users and 80,000+ posts by agents. ## Why did OpenTweet work? ### Timing X turned its API into a meter in February 2026 and asked every developer to run a paid X developer account. OpenTweet launched the same quarter as the layer that hides all of that: connect your account once, post for a flat US$19.99 to US$49 a month. The product's whole pitch is the arithmetic of the meter, which it shows on its own pages. ### One key for an agent and for a person A post an AI agent schedules over the MCP server and a post the user drags onto the calendar are the same record. The agent writes to the queue; the person can review, edit or reschedule it in the app, or revoke the key in one click. That makes the same subscription useful to a developer, an agent and a human marketer. ### Inside X's rules, on purpose The product publishes only through the official X API. It does not scrape, drive a logged-in browser or use the user's password, which are the behaviours that get accounts suspended. It does not auto-reply to strangers, because X blocked that in February 2026, and its DM outreach stops until a human approves each message. The compliance is the selling point against riskier tools. ### Automation on the user's own account Connectors turn RSS, GitHub or Stripe events into drafts, an evergreen queue re-posts the user's best tweets, and auto-retweet re-shares the user's own post hours later to reach a different timezone. All of it acts on the user's own authorized account, so an account can keep posting on a week the owner never opens the app. ## How it was built These details come from the founder-written tech stack on TrustMRR and the company's pages, read on 27 October 2026. They are grade C except the Paddle payment line, which TrustMRR verifies (grade A). - Frontend: Next.js, by the founder-written tech stack. - Backend: MongoDB for data, Auth0 for sign-in, and Docker. - Payments: Paddle, which acts as seller of record for sales tax and which TrustMRR reads for the verified revenue. - AI writing: seven models across Anthropic, OpenAI, Google and xAI, named on the pricing page (Claude, GPT and Gemini models). - A hosted MCP server at mcp.opentweet.io with 43 tools, plus an npm package for local use and a CLI. - Connectors to RSS, GitHub, Stripe and any webhook turn outside events into post drafts. - Analytics through a connected Google Analytics account, by the TrustMRR verification section. Sources: [1, 3, 4] ## Who tried this before OpenTweet, and why did they fail? Four companies in the Loot Drop database built tools on top of social-media platforms and their APIs, and failed when the platform changed its rules, closed its API or built the feature itself. The values below are from that database and are not independently verified. The failed startups come from Loot Drop (https://www.loot-drop.io/), a free database of dead startups, retold here in our own words; Loot Drop describes its entries as AI-assisted summaries of public sources, so they carry no evidence grade. ### Pixelpipe (2008 to 2013; funding US$2M; cause: Competition) - Promise: A single place to post photos and video to many social networks at once. - Why it died: The social networks closed their APIs and added their own cross-posting, so the one-stop tool had nothing left to do. - Record: https://www.loot-drop.io/startup/435-pixelpipe ### Twitter Fabric (2014 to 2020; funding US$100M; cause: Competition) - Promise: A toolkit for app developers, from crash reports to analytics, run under Twitter. - Why it died: Twitter sold it to Google, which folded it into Firebase, so the tool depended on an owner that lost interest. - Record: https://www.loot-drop.io/startup/676-twitter-fabric ### Singly (2011 to 2016; funding US$2.7M; cause: Competition) - Promise: One API that tamed a developer's social data from many networks. - Why it died: It depended on third-party APIs that kept changing, so the single clean interface it sold kept breaking. - Record: https://www.loot-drop.io/startup/548-singly ### Klout (2008 to 2017; funding US$40M; cause: Competition) - Promise: A 1-to-100 score of a person's online influence, built from their tweets and posts. - Why it died: Twitter and Facebook built better analytics of their own, and access to their data for an outside scorer narrowed. - Record: https://www.loot-drop.io/startup/280-klout ## What changed Pixelpipe, Twitter Fabric, Singly and Klout all lived on top of social platforms. Each died the same way: the platform closed or changed the API, built the feature itself, or lost interest in the tool. A product that exists only to make another company's platform easier to use holds its business only as long as that platform allows it. OpenTweet is the same kind of bet, and it says so by building its whole pitch on one platform change: X's move to a metered API in February 2026. For now the change helps it, because a flat monthly price beats a per-post meter, and OpenTweet uses the official API rather than the browser tricks that get accounts banned. But the same company that created the opening can close it: X could change its API prices again, cut off the one-click connection OpenTweet relies on, or offer its own flat plan. TrustMRR's Paddle data shows US$4,025 of recurring revenue on 27 October 2026, earned inside a window that X controls. ## Can one person copy OpenTweet now? **Verdict: Yes, if you accept the platform risk.** The method is clear and needs little money: find a painful change a large platform just made, and sell the layer that hides it, through an API and a calendar, for a flat price. The founder caught X's API meter in the same quarter it appeared. The danger is that the business sits entirely on one company's platform, which can take the opening away. 1. Watch for a platform that just made its API harder or more expensive to use. 2. Sell a flat-price layer on top, with one-click connection and no developer account. 3. Use the platform's official API only, so the accounts you post from are not at risk. 4. Make the same subscription serve an AI agent, a script and a human, so one account covers several buyers. **What is still open.** The same flat-price layer for other metered or hard-to-use APIs: posting to other networks, or wrapping any platform that has just raised its API prices. ## What could go wrong for OpenTweet? - **The business lives on X's API.** OpenTweet exists because X changed its API, and it posts entirely through that API. The same company can change its prices again, end the one-click account connection, or launch its own flat plan. Every predecessor on this page died when the platform it depended on changed the rules. - **Revenue flattened over the summer.** Verified monthly revenue rose fast from US$207 in February to US$3,700 in May 2026, then sat between about US$3,200 and US$4,400 from June to September. The MRR of US$4,025 is at its highest and rose 7.3% in the last 30 days, but the fast early growth has slowed, and October to the 27th shows US$3,455. - **What the product does and does not automate.** The pages read show the product posting, scheduling, cross-posting and auto-retweeting the user's own content through the official X API, and explicitly refusing to auto-reply to other people ("No, and no other tool on a self-serve tier can either") or to drive a logged-in browser. Its DM outreach drafts messages to people who already engaged and "stops" until the user approves each one. Some pages (the dashboard, the /replies, /cross-posting and /calendar paths) are disallowed to ClaudeBot in robots.txt and were not fetched. - **The founder is known by a first name only.** The TrustMRR profile names the founder as "Branko" with the handle brankopetric00 and 20,560 X followers. The sources read do not give a full legal name or confirm who else, if anyone, works on the product. - **User counts are not payment-verified.** The site claims "14,000+ users" and "80,000+ posts by agents", and TrustMRR's founder-written estimate is 14,000 users. Only the revenue is verified: 231 paying subscriptions. The reviews on the home page ("5.0 from 7 reviews") are attributed to named X accounts. - **Always-on trial and a price gap.** The TrustMRR founder-written field still says "from $11.99 a month", while the live site shows US$19.99 as the lowest plan, so the profile text is out of date on price. Every plan requires a card for the 7-day trial, with a 7-day money-back guarantee. ## Sources 1. TrustMRR: OpenTweet, verified revenue profile — https://trustmrr.com/startup/opentweet.md (grade A). Read the markdown profile, downloaded on 27 October 2026 (synced 27 October 2026). Source for the MRR, subscriptions, revenue table and the Paddle verification; the description, team, funding, founded date and user estimate are founder-written (grade C). 2. OpenTweet home page, read 27 October 2026 — https://opentweet.io/ (grade C). Read in full, including the FAQ. Source for the product description, the pricing, the X API meter, the compliance statements, the DM outreach and the AI models. 3. OpenTweet pricing page, read 27 October 2026 — https://opentweet.io/pricing (grade C). Read in full. Source for the three plans, the per-month link-post limits and the annual saving. 4. OpenTweet Replies feature page, read 27 October 2026 — https://opentweet.io/features/replies (grade C). Read in full. Source for the statement that replies are drafted but not auto-sent and that they never touch the X API, because X blocked programmatic replies on 23 February 2026. 5. OpenTweet Auto-Retweet feature page, read 27 October 2026 — https://opentweet.io/features/auto-retweet (grade C). Read in full. Source for the auto-retweet feature, which re-shares the user's own post after a set delay. 6. robots.txt for opentweet.io, read 27 October 2026 — https://opentweet.io/robots.txt (grade C). Read in full. Records that ClaudeBot may read the site but not the /dashboard, /replies, /cross-posting, /ai-prompts, /best-time, /calendar and /api paths, which were not fetched. 7. Loot Drop database export, 1,749 failed startups — https://www.loot-drop.io/ (grade C). Local export. AI-written summaries, not independently verified. Source for the failed predecessors.