SeaWeb MCP
One config block and every MCP-capable agent gets the tools. No SDK.
The config
{
"mcpServers": {
"SeaWeb": {
"url": "https://api.seaweb.tech/mcp",
"headers": { "Authorization": "Bearer sw_your-key-here" }
}
}
}
Where it goes:
| Client | Where the config goes |
|---|---|
| Cursor | .cursor/mcp.json (url + Bearer; manual only — see repo docs/marketplace/cursor-install.md) |
| Claude Code | plugin / claude mcp add — see docs/marketplace/claude-install.md |
| Codex | ~/.codex/config.toml + bearer_token_env_var — see docs/marketplace/codex-install.md |
| VS Code | .vscode/mcp.json |
| OpenCode | opencode.json |
| Windsurf | ~/.codeium/windsurf/mcp_config.json |
Try without a key (SEAWEB_ANON)
Point Claude / ChatGPT custom connectors at:
https://api.seaweb.tech/mcp
Leave auth empty. The anonymous tier serves read-only tools
(search, search_web, get_entity, get_details,
disruption reads, index extract_url) at 20 calls / 60 seconds per client IP.
Writes (remember, …) need a Bearer key. The old
/mcp/k/<key> URL form is retired — keys belong in headers.
Flat meter, not opaque credits: anon is IP-bucketed; keyed callers get per-tool limits
(e.g. search 30/min, extract_url 10/min). Source of truth:
gateway/ratelimit.py.
These apps take a connector URL only — no Authorization header field. Claude Desktop uses the same flow as claude.ai (do not use claude_desktop_config.json for remote URL). Details: repo docs/marketplace/claude-install.md. URL-carried keys were retired 2026-07-30 (HTTP 400 / JSON-RPC -32001). SEAWEB_OAUTH is dark — do not claim keyed Claude.ai private-connector auth.
Tool taxonomy: discover → retrieve → act
| Stage | Tools | Use for |
|---|---|---|
| Discover | list_verticals, search | Pick a vertical; find named entities |
| Retrieve | get_entity, get_details, search_web | Structured fields or crawl passages (read coverage) |
| Act | extract_url, disruption / Watch tools | Pull a cited page; alert on travel disruptions |
Entity-first recipe for agents: paste Search for your coding agent
or the repo SKILL.md. There is no public scrape / map / batch crawl API.
Raw HTTP (no MCP client)
The gateway is plain HTTP JSON-RPC — stateless per request, no handshake, no session:
KEY="sw_..."
curl -s --max-time 20 https://api.seaweb.tech/mcp \
-H "Authorization: Bearer $KEY" \
-H "Content-Type: application/json" \
-H "Accept: application/json, text/event-stream" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
- The
Acceptheader is mandatory.application/jsonalone gets a406. Send both types. - Responses come back as SSE (
event: message/data: {…}). Parse thedata:line as JSON, not the raw body. - Call
tools/calldirectly.initializeis optional and buys a raw client nothing.
Where SeaWeb is listed
Official MCP registry: tech.seaweb/seaweb. Also on Glama. Claude Code plugin: github.com/aritroBh/seaweb-plugin.