A craft · 2 launches
Founders building in AI Agents & Infrastructure
Meet the founders building autonomous agents, model APIs, vector stores, GPU clouds and agent sandboxes, and browse what they have shipped.
Ranked by community votes
-
Number 1: UnoRouter
by Don
Built by a developer tired of juggling AI keys
DR 51 +1 (up) since last check DA 11 no change since last check 2 votes. Sign in to support UnoRouter
Rank = community votes; votes are never sold. DR is Domain Rating by Ahrefs.
The Billboard: paid placement
These founders paid to be introduced here. Whoever has contributed most in total comes first; votes are a different matter.
About AI Agents & Infrastructure
Behind every agent, model endpoint or vector store there is someone who decided to build it. This page gathers those makers and their work: software that carries out tasks on its own, the services that serve language models, the plumbing that keeps them running, the GPU capacity underneath, and the browsers and sandboxes where agents do their work.
Each card puts the founder's face and signature up front, so you can see who stands behind a product before you read the feature list. Newest launches appear first; switch to community votes if you would rather see what others rate, and remember votes cannot be bought. Use the filters to narrow by subcategory, then open a profile to read the founder's own account of what the product does and who it is for.
Browse AI Agents & Infrastructure by type
AI Agents & Infrastructure: questions answered
- What belongs in this category?
- Anything where AI is the product or the foundation: autonomous agents, hosted models and their APIs, tooling for training, deployment and retrieval, GPU capacity for running models, and the browsers and sandboxes agents work in. A tool that only mentions AI in passing is usually better placed in the category matching its real job.
- How do I pick between the five subcategories?
- Ask what the buyer actually does with it. If it completes tasks unattended, choose agents. If developers call it from code, choose LLM APIs and models. If it stores, tunes or monitors, choose MLOps and vector databases. If it supplies compute to run models, choose inference and GPU cloud. If it gives an agent a place to browse or run code, choose agent browsers and sandboxes.
- How do I get listed here?
- Go to /submit and enter your website address. A profile is drafted from your site, and you correct whatever it got wrong before publishing. A paid launch costs $5; a free listing waits in a queue of about 1-2 weeks.
- Can a product sit in more than one place?
- Your listing has a single primary home, chosen at submission and included free. If the product genuinely straddles two areas, the upgrade priced at $5 lets you place it in further categories. Pick the primary one by the main job a buyer hires it for.
- What should I compare when choosing between tools here?
- Read who built it and how they describe the problem, then check how it fits into your existing stack, what you control, and what happens when it fails. Founder notes on profiles are a good place to find limits stated plainly.
