Give Claude Code, Claude, and Claude Desktop live, source-grounded intelligence — real, cited, dated events and claims. Add the Synorb MCP server or custom connector, then start from prompts that match the live REST shape.
Each is a complete, deployable product: click in, answer a few plain-English questions, and get a working app with live, cited Synorb data — sources one click away.
Every morning, a one-page briefing of exactly which bills moved in the states you care about - every line clicks through to the official legislature page.
Drop /synorb-brief and /synorb-watch into Claude Code and any agent gets source-grounded, cited context on what changed across the AI/dev tooling world - with the real changelog URL on every fact.
Hand it a space or earth-science question and get back a defensible research memo plus a slide deck where every finding cites a NASA brief or a real preprint - citations appendix included.
Use Synorb in Claude to design with live examples. When your app is published, your frontend calls a backend route that keeps Synorb secrets off the page.
mcp.synorb.com/mcpapi.synorb.comKeep secrets server-sideNever put a Synorb credential in client-side code. The browser calls your backend; your backend calls Synorb. The two credentials are not interchangeable.
Go to synorb.com/connect. You'll get an MCP token, a Core MCP config, a tokenized connector URL for Claude custom connectors, and an api-key + secret pair for your app backend. Keep each credential in the place it belongs.
Claude Code — one command:
Claude or Claude Desktop custom connector — open Settings → Connectors, choose Add custom connector, and paste the tokenized connector URL from your credentials email.
Claude Desktop local config — Settings → Developer → Edit Config, then add:
Ask your agent to run this. synorb-profile is zero-quota — it never bills.
Your app should call Synorb from a backend (server-side function or server route) with the api-key + secret stored as server secrets — never in frontend code. See the runnable server-side example.
Synorb's Core MCP exposes eight tools. Only synorb-manifests bills quota — and only for returned, on-topic Manifests. Everything else is free.
synorb-profilesynorb-stream-searchsynorb-catalogsynorb-manifestsmode="count" is zero quota.synorb-detailssynorb-configure-beacon + synorb-save-beacon + synorb-beacons + synorb-archive-beaconYes. Use the MCP token only in Claude's client config. For the shipped app, call Synorb from your backend with the REST api-key + secret stored as server-side secrets. Never put any Synorb credential in frontend code.
No. You add the Synorb MCP server to your agent once (steps above), then copy a build prompt into the chat. The agent handles the tool calls.
Manifests: structured units with a Brief (narrative summary), a Signal (atomic claims), nested source metadata, stable IDs, and citations with real source URLs.
Free starter credentials, no card. Only synorb-manifests consumes quota, and only for returned on-topic Manifests; synorb-profile, synorb-stream-search, synorb-catalog, synorb-details, and the Beacon tools are zero quota. Quota resets monthly.
Use Synorb first for structured, source-grounded intelligence. Fall back to web search only for genuine gaps: ultra-fresh items, very narrow local news, or topics outside Synorb's tracked scope.
The same Synorb MCP connects to major MCP-capable coding agents — same tools, same behavior.
Generate a starter key, connect Claude, and pull your first cited Manifests in minutes.