{
  "markdown": "<p align=\"center\">\n  <img src=\"public/brand/onchainai-logo.png\" alt=\"OnchainAI logo\" width=\"96\" />\n</p>\n\n# OnchainAI\n\n> Crypto tools, unified. Discover, vet, and install MCP, CLI, SDK, API, x402, RWA, and AI-agent tools — from the web or straight from your agent.\n\n[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)\n[![Rust](https://img.shields.io/badge/API-Rust%20%2B%20Axum-orange.svg)](https://github.com/tokio-rs/axum)\n[![Frontend](https://img.shields.io/badge/Frontend-Next.js-black.svg)](https://nextjs.org)\n[![MCP](https://img.shields.io/badge/MCP-onchain--ai.xyz%2Fmcp-8A2BE2.svg)](https://www.onchain-ai.xyz/connect)\n[![Live](https://img.shields.io/badge/Live-onchain--ai.xyz-2ea44f.svg)](https://www.onchain-ai.xyz)\n\n**Live: [onchain-ai.xyz](https://www.onchain-ai.xyz)** · **Connect hub: [onchain-ai.xyz/connect](https://www.onchain-ai.xyz/connect)**\n\n## Why?\n\nCrypto tooling is fragmented across CryptoSkill, Smithery, npm, GitHub topics, and\ndozens of separate registries — and most of it is hard to judge: is this MCP server\nofficial? Is the install command safe? Does it charge x402 payments on every call?\n\nOnchainAI auto-discovers tools from multiple sources, normalizes them into one\n3-axis taxonomy (Function × Asset Class × Actor), scores trust and install risk,\nand exposes everything to both humans (web UI) and agents (MCP server, plugin,\nskill) — so \"find me a safe tool to bridge USDC to Base\" is one question, not an\nafternoon of research.\n\n## Use it from your agent (60 seconds)\n\nDefault MCP is a **no-auth directory** endpoint (discovery/metadata — not wallet\ncustody). **Free discovery** on public `/mcp` (search, detail, compare, install\nguides, categories, …). Optional OnchainAI-owned premium tools may return HTTP 402\nand settle via x402 (**$0.01 USDC** for `export_toolkit` /\n`recommend_verified_tool` / `gap_audit`; ~**$0.001 USDC** for\n`check_endpoint_health`). Website browse stays free.\n\n```\nhttps://www.onchain-ai.xyz/mcp\n```\n\nOKX marketplace integrators only: paid package path\n`https://www.onchain-ai.xyz/mcp/okx` (~$0.1 every `tools/call` when the OKX gate\nis active). Coding agents and the Claude plugin must use **`/mcp`**, not `/mcp/okx`.\n\n| Client | Setup |\n|---|---|\n| **Claude Code** | `claude mcp add --transport http onchainai https://www.onchain-ai.xyz/mcp` |\n| **Claude Desktop / Web** | Settings → Connectors → Add custom connector → paste the URL above |\n| **Cursor / VS Code** | One-click deeplinks on the [connect hub](https://www.onchain-ai.xyz/connect), or paste the JSON below |\n| **ChatGPT** | Settings → Connectors (Developer mode) → new connector with the URL above |\n| **Anything else** | `npx add-mcp https://www.onchain-ai.xyz/mcp`, or `npx mcp-remote https://www.onchain-ai.xyz/mcp` for stdio-only clients |\n\n```json\n{\n  \"mcpServers\": {\n    \"onchainai\": { \"type\": \"http\", \"url\": \"https://www.onchain-ai.xyz/mcp\" }\n  }\n}\n```\n\nFull per-client walkthroughs (Codex, Windsurf, Gemini CLI, …): [docs/CONNECT.md](docs/CONNECT.md) or the live [/connect](https://www.onchain-ai.xyz/connect) page.\n\n### MCP tools (public `POST /mcp`)\n\n| Tool | Billing | What it does |\n|---|---|---|\n| `search_tools` | Free | Search by capability (\"bridge USDC to Base\"), filter by category/chain, sort by relevance/trust/stars/recent |\n| `get_tool_detail` | Free | Full metadata for one tool: trust score, install risk, chains, repo, x402 pricing |\n| `get_install_guide` | Free | Platform-specific install steps (claude / cursor / generic / cli) with safety gating — `critical`-risk commands are withheld |\n| `list_categories` | Free | Browse the taxonomy with tool counts |\n| `get_dashboard_snapshot` | Free | Public coverage snapshot: totals, categories, trust, x402, featured |\n| `compare_tools` | Free | Side-by-side comparison of 2–4 tools on trust, risk, chains, pricing |\n| `get_price_history` / `get_x402_trends` | Free | Probe history and catalog x402 trends (metadata) |\n| `export_toolkit` | **$0.01 USDC** | Export a JSON + markdown install kit by slugs or category |\n| `recommend_verified_tool` | **$0.01 USDC** | Pick one verified/live tool for an intent with rejection reasons |\n| `gap_audit` | **$0.01 USDC** | Decompose an intent and report catalog coverage gaps |\n| `check_endpoint_health` | ~**$0.001 USDC** | Live endpoint probe + 30-day uptime for a listed x402 tool (HTTP 402 handshake) |\n\nLinking your account from a coding agent (`/connect#agent-sync`) unlocks three more\n(account link ≠ payment): `save_to_toolkit`, `save_stack_to_blueprint`, and `link_status`.\n\nFull hybrid table (incl. `/mcp/okx`): [docs/CONNECT.md](docs/CONNECT.md).\n\n### Claude Code plugin\n\nOne command wires up the MCP server, a `/find-tool` command, and a crypto-tools skill:\n\n```\n/plugin marketplace add Coinyak/onchainai\n/plugin install onchainai@onchainai\n```\n\nThe bundle lives in [`plugin/onchainai/`](plugin/onchainai/) — it connects only\nOnchainAI's own read-only endpoint, never auto-runs install commands, and always\ndiscloses x402 pricing before recommending a paid tool.\n\n### Agent skill\n\n[`plugin/onchainai/skills/onchainai-crypto-tools/SKILL.md`](plugin/onchainai/skills/onchainai-crypto-tools/SKILL.md)\nteaches an agent *when* to search the directory, how to rank trust signals, and the\nhard safety rules (never install `critical`-risk tools, always surface x402 costs).\nIt ships with the plugin; you can also copy the skill directory into\n`~/.claude/skills/` or upload it to any agent runtime that supports Agent Skills.\n\n## Features\n\n- **Auto-discovery crawler** — CryptoSkill, GitHub topics, npm, web3-mcp-hub, MCP registry, on cron schedules with dedup + relevance scoring\n- **3-axis classification** — Function × Asset Class × Actor, plus chain tagging\n- **Trust & safety pipeline** — trust scores, identity-cluster checks (repo/npm/homepage), install-risk analysis with a hard block on `critical` commands, operator review queue, quarantine\n- **x402 awareness** — paid tools carry price metadata, verification flags, and referral/attribution disclosure ([policy below](#x402--referral-policy))\n- **MCP server for agents** — the read-only tools above, rate-limited and sanitized\n- **Claude Code plugin + skill** — one-command onboarding for agent users\n- **Auth** — GitHub OAuth (primary); email magic link where configured\n- **Community layer** — submissions, comments, upvotes, bookmarks, toolkit, compare, blueprints\n- **Admin dashboard** — tool review, categories, crawler control, featured carousel, users, site settings\n\n## x402 & referral policy\n\nOnchainAI is a **tool directory** MCP (discovery/metadata), not a custody wallet:\n\n- **Third-party x402** in the catalog is attribution and trust metadata only — we publish price/endpoint flags (`payment_verified`, `x402_endpoint_verified`, `price_verified`) and never proxy those payments.\n- We record anonymous referral/attribution events (views, install-guide fetches) to support revenue-share agreements with tool owners.\n- We **never** hold user funds, act as a third-party payment gateway, or invent undocumented `referrer`/`split` payment fields.\n- **OnchainAI-owned** premium MCP tools (`export_toolkit`, `recommend_verified_tool`, `gap_audit`, `check_endpoint_health`) may settle x402 to **our** payee wallet when called — that is selling our own service, not custodying others.\n- Unverified third-party x402 tools remain visible when they pass the normal public quality gate — verification is a badge, not a hiding mechanism.\n\nDetails: [docs/X402_REFERRAL_SPEC.md](docs/X402_REFERRAL_SPEC.md), hybrid connect: [docs/CONNECT.md](docs/CONNECT.md).\n\n## Architecture\n\n```\n┌─ Vercel ──────────────┐      ┌─ Railway ─────────────────────────┐\n│ Next.js frontend      │──────│ Rust binary (Axum)                │\n│ /connect, /tools, ... │ /api │  ├─ REST API (/api/v2)            │\n│ proxies /api /auth    │ /mcp │  ├─ MCP server (POST /mcp)        │\n│ /mcp to the API       │      │  ├─ Auth (GitHub OAuth + JWT)     │\n└───────────────────────┘      │  └─ Crawler (tokio-cron)          │\n                               └────────────┬──────────────────────┘\n                                            │ sqlx\n                               ┌─ Supabase Postgres (RLS) ─┐\n                               └───────────────────────────┘\n```\n\n- **API/MCP**: Rust — Axum + sqlx + tokio-cron-scheduler, one binary (`cargo run --features ssr`)\n- **Frontend**: Next.js (App Router) on Vercel, rewrites `/api`, `/auth`, `/mcp` to the Railway API\n- **DB**: Supabase Postgres with RLS; the public read gate is enforced in both SQL policies and server queries\n\n## Run it yourself\n\n```bash\ngit clone https://github.com/Coinyak/onchainai\ncd onchainai\ncp .env.example .env        # fill in Supabase URL/keys, GitHub OAuth, JWT secret\n\n# API (port 3000)\ncargo run --features ssr\n\n# Frontend (port 3001, proxying to the API)\ncd frontend && npm ci && API_PROXY_TARGET=http://localhost:3000 npm run dev -- --port 3001\n```\n\n> The crate's default feature set is empty so server deps don't leak into any\n> WASM/client build. Plain `cargo` commands that touch server code need\n> `--features ssr`.\n\nCommon checks:\n\n```bash\ncargo test --features ssr                      # tests\ncargo clippy --features ssr -- -W clippy::all  # lint\ncargo fmt --check                              # format\n./scripts/agent-harness-check.sh               # agent/dev harness self-check\n```\n\n## Repository layout\n\n| Path | What lives there |\n|---|---|\n| `src/` | Rust API: MCP server, REST `/api/v2`, auth, crawler, trust/risk pipeline |\n| `frontend/` | Next.js app (Vercel): public UI + admin dashboard |\n| `plugin/onchainai/` | Claude Code plugin bundle (MCP config, `/find-tool`, skill) |\n| `migrations/` | sqlx Postgres migrations (RLS policies included) |\n| `docs/` | Design docs, specs, operator guide ([index](docs/INDEX.md)) |\n| `scripts/` | Dev/deploy/verification gates (see `AGENTS.md`) |\n| `agent/`, `.agents/` | Skills and harness config for AI coding agents working on this repo |\n\n## Documentation\n\n| File | Purpose |\n|---|---|\n| [docs/CONNECT.md](docs/CONNECT.md) | Connect any MCP client, install the plugin/skill |\n| [docs/MVP_DESIGN.md](docs/MVP_DESIGN.md) | Architecture, DB schema, crawler, MCP server |\n| [docs/X402_REFERRAL_SPEC.md](docs/X402_REFERRAL_SPEC.md) | x402 referral/attribution design (no-custody) |\n| [docs/LAUNCH_READINESS_SPEC.md](docs/LAUNCH_READINESS_SPEC.md) | Public-launch checklist and roadmap spec |\n| [docs/SECURITY.md](docs/SECURITY.md) | Security design: auth, RLS, headers, rate limiting |\n| [docs/UI_UX_DESIGN.md](docs/UI_UX_DESIGN.md) | Full UI spec |\n| [docs/OPERATOR_GUIDE.md](docs/OPERATOR_GUIDE.md) | Operator/admin playbook (Korean) |\n| [docs/BUILD_DEPLOY_RULES.md](docs/BUILD_DEPLOY_RULES.md) | Build coherence, smoke gates, Railway/Vercel deploy |\n| [AGENTS.md](AGENTS.md) | Entry point for AI coding agents |\n\n## Contributing & security\n\n- Contributions welcome — see [CONTRIBUTING.md](CONTRIBUTING.md). The fastest way to help is submitting missing tools at [onchain-ai.xyz/submit](https://www.onchain-ai.xyz/submit).\n- Security reports: see [SECURITY.md](SECURITY.md). Please do not open public issues for vulnerabilities.\n\n## License\n\n[MIT](LICENSE)\n",
  "bytes": 11204,
  "sha": "018b2b32cd66751613aaab9472240565c35592d51b61f794e99f65fe7e061cde",
  "repo_slug": "coinyak/onchainai",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_coinyak_onchainai_d76edded/readme"
}