Inspiration
Every AI tool today locks you into a subscription — you pay monthly whether you use it twice or two hundred times. At the same time, developers who build a single useful skill have no easy way to monetize it without building an entire SaaS product around it. We wanted a world where an agent could hire another agent, pay it exactly for the work done, and walk away — the same way you'd hire a freelancer for one task instead of signing them to a full-time contract. That's Algogent.
What it does
Algogent is a marketplace where any skill becomes a live, payable AI agent. A developer uploads a skill definition, and our platform converts it into a fully working agent listed on the marketplace. Users interact through a Chrome extension: connect a wallet, give a task in plain language — apply to jobs, fill a form, automate an email, search the web — and an orchestrator agent finds the right specialist agent in the marketplace, pays it autonomously on-chain, and delivers the result back to the user. No subscriptions, no manual approval for every payment — just pay-per-task, settled directly between agents.
How we built it
The core is an orchestrator agent that breaks a user's request into subtasks and routes each one to the best-matching agent in our marketplace. When a specialist agent is invoked, it responds with a payment-required challenge; our wallet layer signs and settles that payment on-chain automatically, within spending limits the user sets ahead of time. Every payment and interaction is recorded on-chain, so users get a full, transparent breakdown of what was paid, to whom, and for what. Revenue from every paid invocation is split automatically between the agent's developer and the platform.
Challenges we ran into
Getting agent-to-agent payments to feel invisible — no popups, no manual confirmations — while still keeping strict spending guardrails was the hardest part. We also had to design a validation pipeline so that only agents that actually work, tested against real inputs, ever get listed on the marketplace; we didn't want a catalog full of broken or fake agents.
Accomplishments that we're proud of
Our core payment mechanic — autonomous agent-to-agent payments with a policy-enforced spending limit and zero human approval steps — was validated when this idea's foundation won the X402 Global Challenge Prehack in Bengaluru. Algogent is the full platform we built from that result.
What we learned
Building trust into an autonomous payment system is as much a UX problem as a technical one — users need to see exactly where their money went, in real time, or the whole "no human approval" pitch feels risky instead of convenient.
What's next for Algogent
Expanding the marketplace with more real-world agents, opening submissions to outside developers, and building richer spending policies so users can safely delegate bigger, multi-step tasks to their agents.
Log in or sign up for Devpost to join the conversation.