Routers, plugins, dashboards, and agents — shipped by the community on top of the marketplace API. Built something? Get it listed.
Everything the community has shipped on top of the marketplace.
Ranks the order book by quality and supply rather than price alone — layers published intelligence scores and typical pricing over the live listings to show what a benchmark task actually costs to run.
Send a request to the model `auto` and it picks the eligible Surplus model for you, scoring on live prices, a capability registry, and its own routing history. Ships a dashboard, Prometheus metrics, and a feedback endpoint that tunes later picks.
An OpenCode plugin that pulls the whole Surplus catalog into your config so you never hand-maintain a model list — cached models load at startup, then refresh in the background.
An x402Fetch node for the FrostyFi no-code agent builder, with Surplus as a first-class preset. Point a workflow at the API and the agent pays for its own inference — its vault signs the 402, no API key needed.
A Vercel AI SDK provider for Surplus: a thin wrapper over createOpenAICompatible, plus an Anthropic-skin factory and an x402 payment-aware fetch helper.
Benchmarks a set of models on Surplus every ten minutes and tracks each one on JSON-schema reliability, time to first token, tokens per second, and the real per-call cost the marketplace reports back.
A Pi extension that adds Surplus as a model provider — it maps Pi’s thinking level onto reasoning_effort and surfaces reasoning previews from closed models, falling back to token counts when a model reasons silently.
An x402-paywalled LLM gateway. It reads live Surplus prices, sends your request to whichever seller is cheapest for that kind of work, and bills per request in USDC on Base — no account, no API key.
Screen-aware AI in the browser: share a tab and it answers questions about what it sees, watches for changes, and acts for you. The team routes their inference through Surplus.
A unified LLM API for the agent economy: frontier models behind one OpenAI-compatible endpoint, billed from an agent’s own token launch fees or an on-chain balance in USDC, USDT, or any ERC-20.
On-chain investigation in plain English. Ask who is behind a token or whether a wallet is running a pump-and-dump, and it runs the whole forensic workflow across seven chains, with evidence and a confidence label behind every claim.
Smart routing layered on top of Surplus: every call is graded and routed, from a team that had already built routers for Claude Code and OpenClaw.
12 of 12 projects
