# Documentation

- [What is amnt](/docs): An AI SVG generator, an NFT collection launchpad, and a place to publish a paid AI agent. Describe a thing, ship it, get paid.
- [Quickstart](/docs/quickstart): Use an agent in under a minute, or publish one in about five.
- SVG engine
  - [The SVG engine](/docs/svg): Prompt to SVG, or image to SVG. Real editable paths, not a bitmap in an SVG wrapper.
  - [Quickstart](/docs/svg/quickstart): Your first SVG in about sixty seconds.
  - [Prompt to SVG](/docs/svg/prompt-to-svg): What to write, and how the five styles change what comes back.
  - [Image to SVG](/docs/svg/image-to-svg): What converts cleanly, what does not, and how to prepare a file before you send it.
  - [Options](/docs/svg/options): Every field, its type, its default, and what it actually changes.
  - [Export formats](/docs/svg/exports): Seven formats from one document, so a recolour shows up everywhere at once.
  - [The API](/docs/svg/api): Two endpoints, four ways to pay, one response shape.
  - [Pricing](/docs/svg/pricing): What each call costs on each rail, and what costs extra.
- Collections
  - [NFT collections](/docs/collections): One base character, a list of traits, a count. Up to 10,000 SVG pieces, dropped straight to OpenSea.
  - [The bare base](/docs/collections/bare-base): Draw one bald, blank, unclothed body. Everything else is painted onto it.
  - [Traits](/docs/collections/traits): Each trait is painted into one area of the bare body, and lands as its own SVG tile.
  - [Generate and launch](/docs/collections/generate): How many combinations your parts make, how many to generate, and what ships.
  - [Chains](/docs/collections/chains): Base, Ethereum, Robinhood Chain and Hedera. Card to start, wallet optional.
- Concepts
  - [What an agent is](/docs/concepts/agents): A small machine that does one job well, with a page, a price, and an owner.
  - [Connectors](/docs/concepts/connectors): The building blocks an agent is made of — any AI model, any REST API, any MCP tool. Each one is a file, which is why the list keeps growing.
  - [Credits & pricing](/docs/concepts/pricing): One cent per credit, a price floor nobody can undercut into a loss, and a split that survives any cost.
  - [Two doors, one shop](/docs/concepts/two-doors): A person fills a form. A developer sends JSON. A machine pays per call. Same agent, same price, same payout.
  - [Slots](/docs/concepts/slots): The words a buyer can change. The security model and the product model are the same mechanism.
  - [Picture modes](/docs/concepts/picture-modes): Whether the creator's own picture is the style, an example, or both.
  - [Ownership](/docs/concepts/ownership): Every agent has an owner who collects its income. Turning that into a sellable NFT is optional.
- Creators
  - [Overview](/docs/creators/overview): One workshop, one flow, every agent type built the same way.
  - [Writing a recipe](/docs/creators/writing-a-recipe): You are not selling AI image generation. You are selling the words that make it good.
  - [Marking slots](/docs/creators/marking-slots): One glance turns a proven generation into a product.
  - [Picture mode & size](/docs/creators/picture-mode-and-size): One question about how images work, one about what the output is for.
  - [Publishing](/docs/creators/publishing): Mint happens last, not first — you never pay to publish something that doesn't work.
  - [Your agent on OpenSea](/docs/creators/opensea): Every hosted agent is minted on Base and listed on OpenSea's Agent Tool Registry when you publish — free, no wallet, no gas.
  - [The free example on your page](/docs/creators/free-example): Your agent's page shows one of your own past runs, free to read, so a buyer can see what it produces before paying.
  - [Your agent's own site](/docs/creators/agent-sites): Every agent has a web address of its own, chosen by you, with a real website behind it - different for every agent, editable down to the CSS.
  - [Embedding your agent](/docs/creators/embedding): Drop a run card for your agent on any site you already own, in two lines.
  - [Discord Server Builder](/docs/creators/discord-server-builder): An interview-driven full server build - structure, written rules, protection, branding and tickets - run from a server's own Discord.
  - [Discord Ticket Panels](/docs/creators/discord-tickets): Configurable support-ticket panels for Discord - multiple panels per server, each with its own staff, questions, and cleanup rules.
  - [Discord Roles, Giveaways, Scheduling and Backup](/docs/creators/discord-server-extras): Self-assign role panels, a verify gate, giveaways, scheduled posts, and server backup/restore - all webhook-only, no bot uptime required.
  - [Pricing & unlock](/docs/creators/pricing-and-unlock): Three ways a buyer can pay you, and the floors that keep every one of them honest.
  - [Categories](/docs/creators/categories): Ten categories, about sixty subcategories, and every one of them fits one of the four picture modes.
  - [Earned traits](/docs/creators/earned-traits): How your agent's on-chain stats actually grow, and why they lag the website on purpose.
- Buyers
  - [Running an agent](/docs/buyers/running-an-agent): Fill two boxes, pay a few cents, get a result — no account required.
  - [Credits](/docs/buyers/credits): One cent each, bought with a card or crypto. Spend instantly, never a wallet popup per run.
  - [Unlock & ownership](/docs/buyers/unlock-and-ownership): Two ways to get more out of an agent than a single run — use it forever, or own its income.
  - [Hedera NFT tools](/docs/buyers/hedera-nft-tools): Seven free agents that read a Hedera NFT collection — floor price, stats, sales, traits, royalty — from the live market and the chain itself.
- Developers
  - [Developers](/docs/developers): One key, one endpoint, every agent on amnt — callable from your code or from Claude and Cursor as MCP tools.
  - [Quickstart](/docs/developers/quickstart): From no key to a working agent call in three minutes. Nothing to install.
  - [Authentication](/docs/developers/authentication): One Bearer key authenticates the REST API and the MCP server. How keys are stored, scoped, rotated and revoked.
  - [Billing and credits](/docs/developers/billing): 1 credit = 1¢. What a run costs, when you are charged, when you are refunded, and how the creator gets paid.
  - REST API
    - [Run an agent](/docs/developers/api/run): POST /api/v1/run — the endpoint you will use most. Parameters, output shapes, status codes, refunds and timeouts.
    - [List agents](/docs/developers/api/list-agents): GET /api/v1/agents — browse and search every live agent. Free to call.
    - [Get one agent](/docs/developers/api/get-agent): GET /api/v1/agents/{handle}/{slug} — an agent's published input contract, field by field.
    - [Check balance](/docs/developers/api/balance): GET /api/v1/balance — the credit balance your API key spends from.
    - [Errors](/docs/developers/api/errors): Every status code, what causes it, whether you were charged, and whether retrying helps.
    - [Rate limits](/docs/developers/api/rate-limits): 60 requests per minute per key, the headers that tell you where you stand, and how to back off correctly.
    - [OpenAPI spec](/docs/developers/api/openapi): The machine-readable OpenAPI 3.1 description of the amnt API. Import it into Postman, or generate a typed client.
  - MCP server
    - [What it is](/docs/developers/mcp): Plug amnt agents into Claude, Cursor or any MCP client as callable tools. One URL, one key, paid from your credit balance.
    - [Connect your client](/docs/developers/mcp/setup): Config blocks for Claude Desktop, Claude Code, Cursor, VS Code and any other MCP client. Copy, paste, restart.
    - [Tools and scopes](/docs/developers/mcp/tools): Which agents appear as tools, how they are named, how the input schema is built, and the 25-tool cap.
    - [Troubleshooting](/docs/developers/mcp/troubleshooting): No tools showing, 401s, tools that vanish, calls that fail — the causes and the fixes.
  - [Code examples](/docs/developers/examples): A typed TypeScript client, a Python helper, batching, a Next.js proxy route, and shell one-liners.
  - [Long runs and retries](/docs/developers/long-runs): Runs are synchronous. How to handle a 90-second image agent, what to retry, and how to avoid paying twice.
  - [Pay per call with x402](/docs/developers/x402): The account-free rail — every agent is also an x402 endpoint that settles USDC on Hedera per request.
  - [Security model](/docs/developers/security-model): Five walls that make the recipe unreachable — not hidden, unreachable.
  - [Changelog](/docs/developers/changelog): What changed in the amnt developer API and MCP server, newest first, with the versioning rules.
- x402 facilitator
  - [x402 facilitator](/docs/facilitator): A free public facilitator that checks and settles x402 payments and pays the network fee.
  - [Quickstart](/docs/facilitator/quickstart): A paid API route on Hedera testnet in about five minutes.
  - [API reference](/docs/facilitator/api): Every endpoint, what it takes and what it returns.
  - [Networks and fee payers](/docs/facilitator/networks): Which chains are hosted, their ids, USDC addresses and who pays the fee.
  - [Self-hosting](/docs/facilitator/self-hosting): Run your own facilitator with Docker, Node or Vercel. Turn a chain on by adding its key.
  - [Troubleshooting](/docs/facilitator/troubleshooting): The errors you'll actually hit, and the fix for each.
- Platform
  - [Money rules](/docs/platform/money-rules): The rules the money model is never allowed to break, stated plainly.
  - [Trust & guarantees](/docs/platform/trust-and-guarantees): What amnt promises, what it deliberately never promises, and how each promise is actually kept.
  - [Discord bot](/docs/platform/discord-bot): Install amnt in any Discord server to run, build, and fork agents without leaving chat.
  - [Roadmap](/docs/platform/roadmap): What is live, what is being built, what comes after, and what will never be built — stated honestly.
- Reference
  - [Every page on amnt](/docs/reference/pages): The whole product in one table — what each page is for, who it is for, and where the settings live.
  - [Glossary](/docs/reference/glossary): Every term used across these docs, in one place.
  - [FAQ](/docs/reference/faq): The technical questions. Errors, credits, rate limits, suspension, and what happens when something fails.