简体中文
EnglishWebsite review
This review may include AI-generated content and is not individually verified. Domain, certificate and other facts come from the observed data on this page.
What is tinyBots.net?
tinyBots.net is the personal project hub of developer Dave Schumaker: a small collection of free, mostly generative-AI and bot experiments, each presented as a standalone web toy rather than as one unified product. The site itself is essentially an index — you pick a project and use it directly.
What you'll find there
- ArtBot — generates AI artwork from your prompts using a distributed, open-source Stable Diffusion cluster (the "AI Horde").
- Real Fake News — you supply headlines; it asks OpenAI's ChatGPT API to write a matching article, then tries to generate a relevant image via AI Horde.
- Mr. RossBot — a Twitter bot that posts a new Bob Ross–style Stable Diffusion landscape every hour.
- Just Triangles — a Twitter bot generating random triangle patterns from colormind.io color palettes, with an on-site generator too.
- Random Sandwiches — a procedurally generated "book" of 1,449.5 sandwich recipes, made for NaNoGenMo 2017, with source code on GitHub.
Who it's for
Think of it as a showcase for people curious about generative art, procedural generation, and bot culture — and as a portfolio for anyone evaluating Schumaker's work. If you want a polished, account-based creative suite, this isn't that; if you want quick, free, occasionally absurd experiments, it fits.
A practical next step: open ArtBot first. It's the most broadly useful entry point, and trying it tells you quickly whether the AI Horde approach — shared, community-run compute — suits your needs or feels too unpredictable for regular use.
How can I create AI-generated art with ArtBot?
ArtBot is a free browser-based service on tinyBots.net that turns a text prompt into a generated image. It does not run its own AI model; instead, it submits your prompt to the "AI Horde," a distributed, open-sourced Stable Diffusion cluster. That means your request joins a shared queue handled by volunteers' machines, so results depend on who is online and how busy the network is.
What you need to do
- Open ArtBot at tinyBots.net.
- Write a descriptive text prompt — subject, style, lighting and mood all help.
- Submit it and wait for a worker to pick up the job and return an image.
Practical expectations
Because generation is crowdsourced rather than run on dedicated hardware, timing is unpredictable. A short, simple prompt may come back quickly; a detailed one, or one submitted at a busy moment, can take considerably longer. Treat it as a queue, not an instant tool.
If you need faster, more consistent results, compare it with a locally installed Stable Diffusion setup (full control, but needs a capable GPU) or a paid hosted generator (predictable speed, usually a subscription). ArtBot's trade-off is clear: no cost and no setup, in exchange for variable wait times and less control over the model.
Getting better images
- Be specific: "a lighthouse at dusk, stormy sea, painterly, warm light" beats "a lighthouse."
- Mention a style or medium if you have one in mind.
- If the first result misses, adjust the wording rather than resubmitting the same prompt repeatedly.
A reasonable first step is to try one simple prompt and one detailed prompt, then judge whether the wait suits your workflow.
How does Real Fake News generate news articles from headlines?
Real Fake News takes a headline you supply and sends it to OpenAI's ChatGPT API, which writes a full news article to match. It then asks the AI Horde—an open-sourced Stable Diffusion distributed cluster—to generate an image relevant to the story. The result is an illustrated, automatically written article built from a single headline.
What each step does
- Your headline is the only input. You provide the headline; the service does the rest.
- ChatGPT writes the body. The headline is passed to OpenAI's ChatGPT API, which generates a corresponding news article.
- AI Horde supplies the image. The service then attempts to generate a relevant image through the AI Horde, the same distributed Stable Diffusion cluster used by the site's ArtBot project.
Practical trade-offs
The pipeline is fully automated, so the article's tone and accuracy depend entirely on the model's output—this is parody-style generation, not reporting. The image step is an attempt, meaning an image may not always arrive or may not closely match the text. Because both steps rely on shared or external services, speed and availability can vary with demand.
For a concrete test, try a deliberately absurd headline and compare the generated article against the image: the gap between them shows how loosely the image step follows the text. If you want to experiment with the underlying image generation on its own, tinyBots.net also hosts ArtBot.
What is Mr. RossBot and how often does it post?
Mr. RossBot is a Twitter bot listed on tinyBots.net that uses Stable Diffusion to generate a new Bob Ross–style landscape painting and tweet it every hour, around the clock. In other words, it doesn’t post on a human schedule; it produces and publishes a fresh AI-generated painting once per hour, every day.
A practical way to think about it: if you follow Mr. RossBot, you’re subscribing to a steady stream of synthetic landscape art in the familiar “happy little trees” spirit — one new image per hour. For a concrete use case, imagine checking the account during a lunch break and seeing a painting generated since you last looked; the hourly cadence means there’s always something new, but the style stays consistent.
If hourly AI landscape paintings sound appealing, look up Mr. RossBot on Twitter and judge whether the output matches your taste for Bob Ross–style scenes. If you want a different kind of generative art, tinyBots.net also points to related projects such as ArtBot and Just Triangles.
How does Just Triangles generate random triangle patterns?
Just Triangles is a Twitter bot that procedurally generates random triangle patterns, using color palettes pulled from colormind.io. In practice, that means two separate ingredients are combined on each run: a palette that supplies the colors, and procedural code that decides how triangles are arranged, sized and colored across the canvas. The page also notes that you can generate random triangles yourself on the site, so the same generator is available outside the Twitter feed.
What this means in use
- Each output is a variation, not a fixed template. Because the pattern is generated procedurally, the bot isn't replaying one saved design; it builds a new arrangement each time it runs.
- The palette is the main constraint. Colors come from colormind.io, so the visual character of a given pattern depends heavily on the palette it was paired with. The geometry is random; the color relationships are not.
- Two ways to encounter it. Following the Twitter bot gives you a steady stream of patterns. Using the on-site generator lets you produce one on demand, which is more useful if you want to see several results quickly rather than wait for posts.
A practical next step
If you want to judge the range of the generator, open the random triangle tool and generate a handful of patterns in one sitting. Look at whether the variation comes mostly from triangle placement, from color choice, or from both. That tells you whether it's better suited to casual visual browsing or to something like grabbing a quick abstract background.
What is the Random Sandwiches project and how was it created?
Random Sandwiches is a procedurally generated "book" of 1,449.5 sandwich recipes, created by Dave Schumaker for NaNoGenMo (National Novel Generation Month) 2017. It is one of several small, playful internet projects hosted on tinyBots.net.
The project is best understood as a text-generation experiment rather than a cookbook. The recipes are assembled by code, which is why the total is a deliberately odd 1,449.5 rather than a round number. The source code is available on GitHub, so anyone curious about the generation logic can inspect or adapt it.
What it is useful for
- Seeing a compact example of algorithmic writing and combinatorial text
- A conversation piece for people interested in NaNoGenMo or generative art
- A starting point for building your own recipe or instruction generator
Who would find it interesting
Writers experimenting with constrained or generated prose, programmers looking for a small project to study, and anyone who enjoys odd internet artifacts. It is less useful if you actually want to cook a specific sandwich; the output is generated, not tested.
If you want to try something similar, pick a simple template with slots for ingredients, methods and quantities, then let randomness fill the slots. The appeal of Random Sandwiches is not culinary accuracy but the surprising combinations that emerge from a small set of rules.
DigitalOcean
User reviews (0)