Picking a model used to be a coin flip. Now there's real data behind it, and most of it is free. These are the outside references we keep open in a tab while building agents, plus a few reads worth your time.
Nobody paid for a spot here. Some of these people linked to us and some didn't; it made no difference to the list. If something stops being useful, it comes off.
A live leaderboard that compares benchmarks, price per million tokens, and context windows across hundreds of foundation models, refreshed daily from provider APIs. Battle Mode puts two models side by side, which is the fastest way to answer "is the expensive one worth it for my job?" Start here before you commit an agent to a model.
Third-party speed, quality, and cost measurements run on the same hardware conditions for every model. Useful when a vendor benchmark and your own experience disagree.
What developers are sending real traffic to this week, broken out by category. Benchmarks tell you what a model scores. This tells you what people trust it with.
A searchable index of AI tools and agent frameworks, sorted by the job you're trying to get done rather than by category buzzword. Useful when you know the task but not the name of the thing that does it. The write-ups are short enough that you can scan a dozen options in one sitting and still come away with an opinion.
A thirteen-way comparison of agent builders that checked the pricing and free plans itself instead of reprinting vendor claims, split into developer, no-code, and enterprise tiers. The section on testing agents before they hit production is the part most roundups skip, and it's the part that saves you a bad month.
A large, browsable catalog of AI agents and the companies building them, kept current enough to be worth checking before you assume a category is empty. Good for the moment when you know the shape of the problem and want to see everything that claims to solve it.
Seventeen automation tools put through real workflows rather than read off a feature page, and it opens with the right example: the lead you copy from a form into a spreadsheet into a CRM, fifty times a week. The useful part is that it separates no-code from low-code from traditional automation before it starts recommending anything, so you can work out which shelf you're shopping on. No ads and no sponsor funnel, which shows in how bluntly it handles the tools that look alike and aren't.
Ten AI tools assessed for people who are the whole business, where an hour lost to invoicing is an hour nobody else is covering. Opens with a comparison table rather than making you read to find it, and is unusually willing to say which tool is redundant if you already pay for another. The closing sections on building a stack and the mistakes freelancers make are the useful half.
Analyst-style research on building a personal agent, with eighteen platforms mapped alongside it. Longer and drier than a listicle, which is the point: it's one of the few write-ups that treats the build-versus-buy question as a real decision rather than a formality.
Five different routes to your own assistant, laid out side by side so you can see what each one costs you in effort before you start. A good sanity check if you're not yet sure which end of the no-code spectrum you belong on.
A survey of workflow automation tools for people who've hit the edge of what Zapier does well. Written by an automation agency, so read it as informed rather than neutral, but it's specific about where each option starts to hurt.
Eleven workflow automation tools ranked with return on investment as the explicit criterion rather than feature count. Useful if you've got to justify the spend to somebody else.
A platform for hiring AI employees to take on recurring business work across sales, marketing, support, operations, and more. Employees can use company knowledge, work with all tools, collaborate with people and other AI employees, and keep work moving without needing to be prompted at every step.
Discover emerging AI startups, vote in weekly competitions, explore founder spotlights, and see which AI products are rising to the top.
A guide for accounting and law practices between five and fifty people, which is the size band every other roundup skips on its way from solo to enterprise. It opens by saying what is wrong with most AI tool guides and then avoids doing those things, sorting picks by the kind of firm you run rather than by feature count.
Eight tools for operations managers, each with pricing, a use case, and a direct vendor link, published with a stated no-affiliate policy. That last part is rarer than it should be, and it is why this list reads like advice instead of a commission schedule.
A comparison of operations tooling that publishes its methodology and a review date instead of asking you to take its word for it. Strongest on the buying-guide section, which walks through what to check before you commit a team to a platform.
A reference guide covering 131 AI tools for business, each with a real explanation rather than a one-line label. Long enough to be a reference rather than a read, and arranged so you can jump to the job you're trying to do.
Seven agents assessed for businesses that do not have an IT department to absorb the setup cost. Independent, and refreshingly willing to say when a tool's aimed at a bigger company than yours.
AI tooling ranked for franchise management, a corner of the market where the operational constraints are genuinely different and almost nobody writes about it carefully. Run by people who work in franchising rather than in software.
Ten AI tools for insurance agencies, sorted by the job rather than the vendor, so an agent can see where the day goes before spending anything. Independent of the carriers and the software makers, which is why the pricing notes read like notes instead of a brochure.
A guidebook for contractors, plumbers, and HVAC shops that opens with the number of calls that go unanswered rather than a robot picture. Covers phone answering, field service platforms, and sales coaching, and it is honest about which of those pay for themselves in a week and which do not.
What a CPA firm should settle before adopting AI in tax work, written from inside the profession rather than from a vendor's marketing team. The sections on review responsibility and client data are the ones that're worth reading twice.
A deliberately short, honest look at AI bookkeeping tools for small businesses, including where the category still doesn't deliver. Short is a feature here.
Ten HR tools covering both free and paid options, reviewed independently rather than as a vendor roundup. Useful if you're a team of one doing HR alongside everything else.
Twelve sourcing tools compared by a publisher that has been writing for developers since long before anyone called this AI. Useful if you're hiring and want the category mapped before you start a trial, rather than after.
Twelve tools for farms, covering the field and the office rather than just the drone footage everyone leads with. One of the better vertical roundups we've come across.
Written by people who work with farm data, which shows in the choice of what to include. Practical about the gap between what an agricultural AI tool promises and what it'll do on a working farm.
Twenty-six tools for working photographers, with the free options separated from the paid ones honestly instead of blurred together. It covers the business and admin side of a photography practice as well as the editing, which is the half most roundups forget.
A long-running mobile photography publication, which means the perspective here is a photographer's rather than a technology reviewer's. Worth reading alongside any tools list written by somebody who doesn't shoot.
Enhancing, retouching, and restoring, kept to a tight set of picks rather than padded to twenty. Independent roundups this tight are rarer than they should be.
Thousands of real marketing emails, saved and searchable by type, so you can see how somebody else already solved the layout you're stuck on. We point people here whenever they ask what a welcome sequence is supposed to look like, because an afternoon of scrolling this beats a week of guessing.
If you're new to this and want a working example rather than a spec sheet, these three are the shortest path from idea to running agent: email automation, meeting notes, and data entry. Or browse every workflow and pick the one that matches your Monday.
Run a tool, tracker, or guide that belongs on this page? Email [email protected] with a line on why it's useful. We read everything and we link to what earns it.