Inbox Zero

A user gives a topic, a text, a PDF, a web address or a YouTube link, and a language model writes a slide presentation from it. MagicSlides runs as an add-on inside Google Slides and as a web application that exports editable PowerPoint, PDF and Google Slides files. Since July 2026 the same account also includes AI tools for spreadsheets, documents, meeting notes and a design canvas, under the name Magic Suite.

It appeared in January 2023 inside Google Slides, a program where people already work, when writing slides from one sentence was new, and users showed it to each other without paid advertising. A blog about presentation topics then brought about 80% of new users from Google search, by the founder's account. Both channels weakened: search traffic fell, and Google added the same function to Slides. The verified total is real, and so is the fall: monthly revenue fell by half between February and July 2026.

A web app that turns a topic, notes, a PDF, a Word file, a spreadsheet, a YouTube link or a web page into a slide deck. The user approves an outline and a visual concept, then each slide is generated as one designed full-slide image, edited by chat, and exported to PowerPoint (.pptx), PDF or PNG.

It is run by a search marketer: its pages target what people type into Google, and its best month came from a list article on another company's blog that ranked there. A US$1 trial that needs a card turns visitors directly into subscribers at US$25 a month or US$99 a year.

You upload a PDF and ask questions about it in a chat window. The answers come from the text of the document. Paid plans add more uploads, more questions, OCR for scanned pages, a Chrome extension, chat across several PDFs, and an API for developers.

The founder did not build it. He bought a weekend project with no revenue for US$20,000, paid about US$10,000 for the domain pdf.ai, and relaunched it under that name. The domain put the product at the top of Google for the words people were already typing, and a one-time launch deal paid for a developer and a video creator within weeks.

A calendar that publishes posts to more than 30 social networks: X, LinkedIn, Instagram, TikTok, YouTube, Reddit, Bluesky, Discord and others. The whole code is open source, so anyone can run it on their own server for free; the company sells the hosted version from US$29 a month. Since 2025 AI agents such as Claude, ChatGPT and OpenClaw can connect to it through an MCP server, a command-line tool and a public API, so an agent drafts and schedules the posts and a person approves them in the calendar.

The founder took a twenty-year-old product category with no open-source entry and open-sourced all of it, which gave him millions of Docker downloads and 37,000 GitHub stars without a marketing budget. Revenue stayed small for a year. It multiplied when he turned the product toward people who automate their posting with n8n workflows and, from late 2025, with AI agents, which need an API-first scheduler and keep paying because they post without human effort.

A web app that designs the screens of an iPhone or Android app from one sentence or a reference image, with AI. You chat to change the design, then paste it into Figma as editable layers, export HTML or React code, or send it to an app builder. AI coding agents can also use it through an API.

It took a narrower job than the general AI design tools, mobile apps only, for people who build apps without design skills. The three founders reused what they had learned from four earlier design tools, built it in three weeks, and showed finished designs on X before the launch. Each new Gemini model gave them a better result and a new demo to post.

A company writes instructions for an AI agent, connects its documents and tools (Notion, Google Drive, HubSpot, Zendesk, Jira), and places the agent in Slack or Microsoft Teams with a name such as @Support or @SalesOps. Team members mention the agent in a channel; it answers from the connected knowledge, drafts replies and briefs, and updates records after a person approves. No code is needed.

The first version was an AI copilot for DevOps engineers, who could build their own agents. The founders moved the product to non-technical teams that answer the same questions all day, started every customer with one small use (repeated questions in Slack), and added a higher-priced plan when power users asked for more agents. The founder said in an interview of 17 October 2025, as quoted by SoFarBot, that this took the company from US$100,000 to US$400,000 in annual recurring revenue.

An AI agent for paid advertising that works inside ChatGPT, Claude, Claude Code, Cursor and other AI assistants. It connects to the customer's ad accounts at Google, Meta, LinkedIn, TikTok, Amazon and ChatGPT Ads, researches keywords, builds campaigns (always created paused), reports results and finds wasted spend. Scheduled agents can also watch the accounts and make reversible changes. The customer pays the ad platforms directly.

It went where the users already were: an app in ChatGPT's app store, a plugin in Claude's directory and in Claude Code, and setups for Cursor and Codex, in the months when these directories were new. It charges per tool call instead of a share of ad spend, so a small advertiser starts at US$49 a month. And it applied to the platforms' own API programs; the company says Google approved it in June 2026 and Amazon admitted it to its partner network.

A LinkedIn lead finder for B2B sales teams. You list LinkedIn profiles to watch; traxy collects the people who like, comment on or repost their posts, scores each one against your ideal buyer, adds an email address and phone number, and an AI agent can write and send the first message.

The founder already ran a LinkedIn ghostwriting agency and built the lead tool that the agency needed. He sold it with the tool itself: traxy found the leads, and his team contacted them. Daily posts on X added trials, and credit-based prices let heavy users pay more.

A web application for people who sell simple books through Amazon's Kindle Direct Publishing. The user describes a book, and AI models produce its pages: line drawings for coloring books, puzzles and mazes for activity books, text and layouts for journals and planners, and a description for the Amazon listing. A subscription buys credits. By the terms of service one generated image costs 200,000 credits, so the smallest plan, with 40 million credits a month, pays for about 200 images.

It sold one narrow job to a group that already existed: people who learn from YouTube channels and paid courses how to publish many simple books on Amazon. The product carries that group's own word in its name, tutorial channels in English and Italian showed it from January 2025, and the site runs an affiliate script. Two thirds of the US$642,409 came in the first 14 months, at about US$30,300 a month. The last 12 months average about US$17,600 a month.

One API address, in the same format as OpenAI's API, that forwards each request to one of more than 200 models from more than 40 providers, such as OpenAI, Anthropic and Google, and records the cost of each request. Customers prepay credits at the providers' prices plus a 5% fee, bring their own provider keys for free, or run the open-source code on their own servers. DevPass is a monthly subscription for AI coding agents.

Every team that builds with AI models ends up with several provider accounts, keys, invoices and SDKs. One gateway with open code replaces them, and a developer can read the code before sending traffic through the hosted version. In 2026 the founders added DevPass, a flat monthly price for coding agents with an allowance of up to three times the price in model usage. The verified monthly receipts grew from US$57,994 in January 2026 to US$249,922 in August 2026; TrustMRR does not split them by product.

A web app that turns a screenshot, a mockup, a Figma design or a screen recording of a website into front-end code: HTML with Tailwind or CSS, React, Vue, Bootstrap or Ionic. The code is open source under the MIT licence and free to run with your own model keys. The hosted version at screenshottocode.com sells monthly credits, so the user needs no API key and no setup.

The founder released it as a free open-source tool nine days after he wrote that the new GPT-4 Vision and Turbo API previews allowed only 100 requests a day. Hacker News and GitHub trending brought the repository more than 35,000 stars in under three months, and the README sends readers on to the paid hosted version for people who do not want to set up keys and servers. It had made US$1,000 within 19 days, with a Stripe payment link and no user accounts.

A set of free Python packages for designing optical parts with AI methods: a gym of standard design problems (invrs-gym), optimization algorithms (invrs-opt), a light simulator that runs on GPUs (fmmax), and a public leaderboard of the best solutions. Everything is MIT-licensed code on GitHub, installed with pip. There was no paid product.

One researcher who had led AI-guided optics design at Alphabet and Meta put the tooling of that field into the open, with a shared benchmark that anyone could run. Apple, which designs cameras, displays, sensors and lidar parts, bought the assets and hired him. The buyer's filing with the European Commission is the only hard document; it gives no price.