{
"markdown": "\n\n# Prism — Crossdeck's intelligence layer for AI\n\n**Ask your app anything — revenue, errors, database read-cost, growth — in plain English, and get rendered charts and dashboards back.**\n\n[](https://www.npmjs.com/package/@cross-deck/ai)\n[](./LICENSE)\n[](https://modelcontextprotocol.io)\n[](#security)\n[](#tools)\n\n*\"Which paying customer did this crash hit, and how much revenue is at risk?\" — one question, answered across error, identity, and revenue. No single-layer tool can do that.*\n\n[**What Prism does →**](https://cross-deck.com/prism/) · [Developer docs](https://cross-deck.com/docs/prism/) · [Crossdeck](https://cross-deck.com)\n\n
\n\n---\n\n## Why Prism is different\n\nMost MCP servers are a connector to *one* system. **Prism is an intelligence layer over your whole app** — because Crossdeck already joins the layers about your users (**identity · revenue · entitlements · errors · analytics · database read-cost**) into one source of truth, *by identity*.\n\nThe point isn't \"let AI read my data.\" It's **the crossing**: one question can span layers no other tool joins —\n\n> *\"This error `a1b2c3` — who did it hit, and how many of them pay us?\"*\n> → the named paying users behind the crash, each with their monthly revenue at risk, sorted by what's on the line.\n\nThat's the moat sentence. Prism gives it a voice in Claude, ChatGPT, Cursor, or any MCP client.\n\n---\n\n## Permissions & data access\n\n**Secure, read-only access to your live Crossdeck workspace.** Prism reads; it never writes.\n\n**Prism can see** (the layers Crossdeck already owns about your app):\n\n| | |\n|---|---|\n| 💷 **Revenue** | MRR, paying-customer counts, per-rail split |\n| 🐛 **Errors** | issues, blast radius, who was affected |\n| 📈 **Analytics** | per-host page views, unique visitors, top pages/referrers |\n| 🗄️ **Read-cost** | per-user-vs-overhead database reads, by operation |\n| 🧭 **Customer journeys** | the cross-layer view of one customer (pay × entitlements × cost) |\n\nRow-level **customer identity** is returned only when your connection was explicitly granted that scope — otherwise Prism answers from aggregates and tells you so. That boundary is enforced by Crossdeck, not left to the model.\n\n**Prism cannot** create, modify, delete, or configure anything in your projects. There is no write path — every tool is read-only (`readOnlyHint: true`). Ask it to change something and it will tell you it can't.\n\nData is fetched live at question time over an OAuth-protected endpoint; the connector stores nothing.\n\n---\n\n## Connect in 60 seconds\n\n### Remote (recommended) — one-click OAuth, no keys to paste\n\nPrism is a hosted, OAuth-protected MCP endpoint:\n\n```\nhttps://api.cross-deck.com/mcp\n```\n\nAdd it as a connector in your client, sign in with your Crossdeck account, pick a project, and approve. That's it — the client receives a short-lived, scoped workspace credential automatically.\n\n- **Claude (web / desktop):** Settings → Connectors → Add custom connector → paste the URL above.\n- **ChatGPT:** Settings → Connectors → Add → paste the URL.\n- **Cursor:** Settings → MCP → Add → paste the URL.\n\nOr open your dashboard at **app.cross-deck.com → Developers → Prism → Connect AI** for one-click \"Add to Claude / ChatGPT / Cursor\".\n\n### Local (Claude Desktop / Cursor / Claude Code) — stdio + a secret key\n\nNo install step — `npx` fetches the package **and its dependencies** on each launch, so it\nself-heals and stays current (nothing local to go stale):\n\n```json\n{\n \"mcpServers\": {\n \"crossdeck\": {\n \"command\": \"npx\",\n \"args\": [\"-y\", \"@cross-deck/ai\"],\n \"env\": { \"CROSSDECK_SECRET_KEY\": \"cd_sk_live_…\" }\n }\n }\n}\n```\n\nUse a **secret** key (`cd_sk_`) from your dashboard → API keys — never a publishable (`cd_pub_`) one. `CROSSDECK_API_BASE` overrides the endpoint (e.g. sandbox).\n\n> **If the connector shows \"failed\" / \"Server disconnected\" — two known causes, both fixable:**\n>\n> 1. **`spawn npx ENOENT` / command not found.** Desktop clients launch with a *minimal*\n> PATH that often excludes your Node bin (common with `nvm`). Fix: use the **absolute\n> path** to `npx` as `command` — run `which npx` and paste the full path (e.g.\n> `/Users/you/.nvm/versions/node/vX/bin/npx`).\n> 2. **`ERR_MODULE_NOT_FOUND … @modelcontextprotocol/sdk`.** This only happens if `args`\n> points at a **hand-built local checkout** (`…/dist/server.js`) — that folder's\n> `node_modules` can be cleaned out from under you (e.g. anything under `/tmp`). Fix:\n> **never point at a local build.** Let `npx -y @cross-deck/ai` (above) resolve the\n> published package, so its dependencies are always present.\n>\n> Simplest of all: use the **remote connector** at the top of this section — it runs no\n> local Node at all, so neither failure can occur.\n\n---\n\n## Try these first\n\n| Ask | Tool | You get |\n|---|---|---|\n| *\"What's our MRR and paying-customer count, split by Stripe/Apple/Google?\"* | `get_revenue` | MRR, paying count, three-rail split, optional daily trend |\n| *\"This error `a1b2c3` — who did it hit and do any of them pay us?\"* | `get_error_affected_users` | The **named paying users** behind the crash, each with monthly revenue at risk |\n| *\"Show me everything about customer `agent_8842`.\"* | `open_moat_dashboard` | A **rendered dashboard**: what they pay × entitlements × read-cost, joined by identity |\n| *\"Draw our user growth over the last 30 days.\"* | `draw_user_growth` | A **rendered interactive line chart** of visitors + page views |\n| *\"What's driving our database reads — per-user vs overhead?\"* | `get_read_cost` | The per-user-vs-overhead split + reads by operation |\n\nData tools return clean JSON; `draw_user_growth` and `open_moat_dashboard` render interactive charts/dashboards inline in hosts that support MCP Apps (and fall back to a text summary elsewhere).\n\n---\n\n## Tools\n\nThirteen tools — all **read-only** (`readOnlyHint`), each with a human-readable `title`, scoped/paginated output, and actionable errors. Every read is a point-read of a maintained ledger, so **asking questions never runs up your database bill.**\n\nEvery number is **self-describing**: values come back with a `meta.semantics` briefing and a `coverage` state, so a metric is reported as exactly what it is, and a blind spot (`not_instrumented`) is never mistaken for a real zero.\n\n**Portfolio**\n- `list_projects` — list every app this connector can read (your portfolio).\n- `get_portfolio` — the coverage map: for each app, which surfaces are instrumented vs a blind spot. Load it first so a later blank reads as a known blind spot, not a fact.\n- `use_project` — set the current app for subsequent tools (\"switch to \").\n\n**Revenue & cost**\n- `get_revenue` — MRR, paying-customer count, per-rail split (Stripe/Apple/Google), optional daily trend.\n- `get_customers` — the customer census: the paying-customer count (plus total / active where instrumented). The canonical \"how many customers does this app have\" answer.\n- `get_read_cost` — database read-cost split into per-user reads vs un-attributed overhead, by operation.\n\n**The moat — error × identity × revenue**\n- `get_error_impact` — for one error: how many distinct users hit it, and how many are paying (counts).\n- `get_error_affected_users` — the **named paying users** an error hit, each with monthly revenue and last-hit, sorted by revenue at risk. *Returns your own identifiers only — never emails/names you didn't provide.*\n- `get_customer` — one customer across every layer Crossdeck joins by identity: what they pay × active entitlements × read-cost.\n\n**Analytics**\n- `get_host_analytics` — page views + unique visitors for a host/subdomain you own (optional daily series).\n- `get_host_top_pages` — top pages or referrers for a host you own.\n\n**Rendered UI (MCP Apps)**\n- `draw_user_growth` — an interactive line chart of visitors + page views over time.\n- `open_moat_dashboard` — a customer's cross-layer dashboard (pay × entitlements × read-cost) rendered as cards.\n\n---\n\n## Security\n\nPrism is built on Crossdeck's bank-grade outbound-read spine. See [SECURITY.md](./SECURITY.md) for the full model. In short:\n\n- **OAuth 2.1 + PKCE (S256)**, dynamic client registration, short-lived (1h) workspace credentials with refresh-token rotation and reuse detection. No long-lived secret leaves your dashboard.\n- **Read-only.** Every tool is `readOnlyHint: true` — no writes, no config mutations.\n- **Scoped & fail-closed.** A token resolves to one owner and their own projects; a missing/invalid token returns an actionable 401, never silent data.\n- **Your data only.** Prism reads your own project's data, scoped by your token. It does **not** read your conversation history or local files, and returns your own identifiers — never PII you didn't provide.\n\n---\n\n## Known limitations\n\n- **Read-only** in v1 — writes/config are a later, separately-gated stage.\n- Per-host analytics require the host to be a **verified origin** of your project.\n- Rendered charts require an MCP-Apps-capable host; others get a text summary.\n- Reporting reflects data from when Crossdeck was connected (historical backfill is separate).\n\n---\n\n## Links\n\n- **npm:** https://www.npmjs.com/package/@cross-deck/ai\n- **API reference:** https://cross-deck.com/docs/reporting-api/\n- **Security spine:** https://cross-deck.com/docs/outbound-api-security/\n- **Privacy:** https://cross-deck.com/legal/privacy/\n- **Support:** support@cross-deck.com\n\n## License\n\n[MIT](./LICENSE) © Crossdeck\n",
"bytes": 9856,
"sha": "424a96ef9c6a8dc2aa2090b77457e2da490a168f965d442906ddaede4b83b1c2",
"repo_slug": "crossdeckhq/crossdeck-prism",
"fonte": "repo",
"truncated": false,
"api": "https://agentalog.com/api/listings/mcp_com_cross_deck_prism_9ed8ca7a/readme"
}