whoami
Who am I
Identify the API key this connection uses (GET /me): `tenantId` (the workspace), `apiKeyId` and `scopes`. Write tools need the `write` scope; call this to check before attempting them.
Takes no parameters.

io.github.gege-mnPublished by GitHub account gege-mnv0.1.0Updated Oct 24, 2026
Send email, push, Slack, Discord, Telegram and webhook notifications and manage templates with Ray.
Ray's notification delivery MCP server: lets an assistant send email, push, Slack, Discord, Telegram and webhook notifications and manage templates, delivery, a
Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.
Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.
Read https://sourceweft.com/mcp/io.github.gege-mn%2Fray-mcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.Add this to your client's mcpServers config.
{
"mcpServers": {
"ray-notifications": {
"type": "http",
"url": "https://ray-api.gege.mn/mcp",
"headers": {
"Authorization": "Bearer <YOUR_TOKEN>"
}
}
}
}MCP server for Ray, the notification delivery API. It lets Claude, Cursor, VS Code, Windsurf, Codex, Zed and any other MCP client send notifications and manage templates, delivery webhooks and usage in your Ray workspace.
Ray here is the multi-tenant notification API from gege.mn (email via Amazon SES or SMTP, Firebase push, Slack, Discord, Telegram, SMS via Twilio or sendsms.mn, generic webhooks and an in-app feed). It is not the Ray distributed computing framework.
Ray runs an MCP server for you. There are two ways to connect, with the same tools:
This package is a thin bridge: it connects to the hosted endpoint with your key and mirrors its tool list, descriptions and instructions, so new tools appear without upgrading the package. All validation, scopes, rate limits and quota are enforced by Ray's API.
ck_live_ and is shown only once.
read scope: status, templates, usage, docs.write scope: also needed to send, create or change templates, and manage webhooks.The examples below use ck_live_... as a placeholder. Keep real keys out of version control.
Hosted:
Local:
Add --scope user to make it available in every project. Check the connection with claude mcp get ray.
Edit claude_desktop_config.json (Settings → Developer → Edit Config; macOS ~/Library/Application Support/Claude/claude_desktop_config.json, Windows %APPDATA%\Claude\claude_desktop_config.json), then fully restart Claude Desktop:
.cursor/mcp.json in a project, or ~/.cursor/mcp.json for all projects.
Hosted:
Local:
Cursor can read the key from your environment instead: "Authorization": "Bearer ${env:RAY_API_KEY}".
VS Code's mcp.json (.vscode/mcp.json in a workspace, or MCP: Open User Configuration) uses a top-level servers key and can prompt for the key so it isn't stored in the file:
Local variant: replace the server entry with { "type": "stdio", "command": "npx", "args": ["-y", "@gege-mn/ray-mcp"], "env": { "RAY_API_KEY": "${input:ray-api-key}" } }.
~/.codeium/windsurf/mcp_config.json (Windsurf calls the URL field serverUrl and supports ${env:VAR}):
Local: use the same command / args / env entry as in the Cursor example.
~/.codex/config.toml, local:
Hosted, reading the key from your shell environment:
Or from the CLI: codex mcp add ray --env RAY_API_KEY=ck_live_... -- npx -y @gege-mn/ray-mcp.
In Zed's settings.json:
Local: "ray": { "command": "npx", "args": ["-y", "@gege-mn/ray-mcp"], "env": { "RAY_API_KEY": "ck_live_..." } }.
Any client that runs stdio servers:
Any client that supports Streamable HTTP: URL https://ray-api.gege.mn/mcp, header Authorization: Bearer ck_live_....
CLI flags: --help, --version. Requires Node.js 18 or newer.
The tool list comes from the hosted server, so it can grow over time; your client shows the authoritative descriptions and input schemas. Current tools:
Read-only tools are annotated readOnlyHint; archive and delete tools are annotated destructiveHint, so clients can ask before running them.
firstName param, then test-send it to me."snd_... get delivered? Show me any failures."https://example.com/hooks/ray for failed deliveries."The local package always starts, even when something is wrong, and explains the problem through MCP instead of crashing. Diagnostics go to stderr (your client's MCP log); stdout carries only MCP messages.
setup_help tool is listed. The server can't load Ray's tools. Its description says why. Without RAY_API_KEY every tool call returns setup steps. If the endpoint was unreachable, call setup_help after fixing the cause: it retries, and on success the client is told the tool list changed. If your client doesn't refresh, reload the server.write scope.get_usage.RAY_MCP_URL.npx not found or old Node. Install Node.js 18+ and make sure npx is on the PATH your client uses. On macOS GUI apps may not see shell PATH changes; use an absolute path to npx if needed.RAY_API_KEY=ck_live_... npx -y @gege-mn/ray-mcp should log connected to https://ray-api.gege.mn/mcp on stderr, then wait for input (Ctrl+C to exit). Or inspect it with npx @modelcontextprotocol/inspector npx -y @gege-mn/ray-mcp..md to any docs URL for raw markdown). The MCP setup guide is Using Ray with AI coding agents.@gege-mn/ray (gege-mn/ray-node).npx skills add gege-mn/ray-skills).Releasing: bump the version in package.json, server.json (both version fields) and src/version.ts, update CHANGELOG.md, then publish a GitHub release tagged vX.Y.Z. The publish workflow ships to npm and the MCP Registry. To publish by hand instead: pnpm publish --access public, then run the MCP Registry workflow (a personal mcp-publisher login github can't publish under the gege-mn namespace).
MIT © gege.mn
Source: README.md at commit ebdfa1b
Who am I
Identify the API key this connection uses (GET /me): `tenantId` (the workspace), `apiKeyId` and `scopes`. Write tools need the `write` scope; call this to check before attempting them.
Takes no parameters.
Get plan and usage
The workspace's plan, subscription status and this calendar month's usage against the monthly quota (GET /billing/subscription). `used`/`remaining` are null on unmetered plans. On a hard-capped plan, sends beyond the quota fail with HTTP 402 quota_exceeded.
Takes no parameters.
List channels
List the delivery channels configured in this workspace (GET /channels). Call this FIRST before sending: each entry's `id` is the `channelConfigId` for send_notification and test_send_template, `kind` is the channel type (ses_email, smtp_email, fcm_push, slack_webhook, discord_webhook, telegram_bot, twilio_sms, sendsms_mn, generic_webhook), `templateKind` is the template channelKind it accepts, and `recipientSchema` is the exact JSON Schema of the `recipient` it expects. Channels and their provider credentials are set up in the Ray dashboard, not through this API; if the list is empty, ask the user to add one there.
Takes no parameters.
Send notification
Send a notification (POST /send; needs a `write`-scoped key). Returns `{ sendId }` with HTTP 202: delivery is asynchronous, so check the outcome with get_send_status. Use exactly one mode: 1. Single: `channelConfigId` + `recipient`. 2. Fan-out: `channelConfigId` + `targets` (1-1000 `{ recipient, externalUserId? }`), one channel, one sendId. 3. Multi-channel: `deliveries` (1-10, each with its own `channelConfigId`, `recipient` and `templateId` or `content`) for ONE person over several channels; add `feed` + `externalUserId` for a single in-app feed entry. 4. Feed-only: `feed` (with `title`) + `externalUserId` and no channel: an in-app notification only. In modes 1-2 give exactly one of `templateId` (a published template, plus `params`) or inline `content`; in mode 3 that choice is made per delivery. Get channel ids and each recipient shape from list_channels first. Pass `idempotencyKey` whenever a retry must not deliver twice.
feedparamscontenttargetslogTitleprioritynotBeforerecipientcampaignIddeliveriesshowInFeedtemplateId+5 moreGet send status
Delivery status of a send (GET /sends/{id}): `aggregate` counts per status over the whole send (pending, claimed, delivered, failed_retryable, failed_terminal, suppressed) plus one page of per-recipient delivery rows including provider errors. Right after sending, rows are usually `pending`; check again after a few seconds. Feed-only sends have no rows. For more rows, pass the returned `nextCursor` as `cursor`.
limitcursorsendIdrequiredList templates
List message templates (GET /templates): id, name, folder, channelKind, and `publishedVersionId` (null = never published, so it cannot be sent by templateId yet). Archived templates are hidden unless `includeArchived` is true. Use get_template for content and required params.
folderincludeArchivedGet template
Get one template with its published version and its current draft (GET /templates/{id}): content, logTitle/logDescription, `requiredParams` a send must supply, and paramOverrides. Read this before update_template_draft, which needs the full body.
templateIdrequiredCreate template
Create a template (POST /templates; write scope). `channelKind` is permanent and must match the channels you'll send through (list_channels `templateKind`). It is saved as an unpublished draft unless `publish: true`; sending by templateId needs a published version (publish_template). Names are unique per workspace including archived templates: a collision returns 409, so unarchive_template the old one instead. Returns `{ id, requiredParams }`. Designed (drag-and-drop) email templates can only be made in the dashboard.
namerequiredfoldercontentrequiredpublishlogTitlerequiredchannelKindrequiredlogDescriptionrequiredparamOverridesUpdate template draft
Replace a template's draft (PATCH /templates/{id}; write scope). This is a full replacement, not a partial patch: pass content, logTitle, logDescription and paramOverrides as they should end up (get_template first). `name`, `folder` and `channelKind` are required for validation but not changed (channelKind must equal the original). Live sends keep using the published version until you publish, either with `publish: true` here or publish_template. Fails on archived templates.
namerequiredfoldercontentrequiredpublishlogTitlerequiredtemplateIdrequiredchannelKindrequiredlogDescriptionrequiredparamOverridesPublish template
Publish a template's current draft as its new live version (POST /templates/{id}/publish; write scope). Sends by templateId use it immediately. Fails if there is no draft or the template is archived.
templateIdrequiredArchive template
Archive a template (POST /templates/{id}/archive; write scope). Sends that reference it fail until it is restored with unarchive_template; its name stays reserved meanwhile. Confirm with the user first if it may be in production use.
templateIdrequiredUnarchive template
Restore an archived template (POST /templates/{id}/unarchive; write scope). This is how to reclaim a name held by an archived template instead of creating a new one.
templateIdrequiredTest-send template
Send a real test of a template's PUBLISHED version to one recipient (POST /templates/{id}/test-send; write scope). It goes through the provider for real, so use a recipient the user controls. Test sends don't count toward the monthly quota (a small daily allowance applies instead) and never appear in feeds. Returns `{ sendId }` for get_send_status.
paramsrecipientrequiredtemplateIdrequiredchannelConfigIdrequiredList feed notifications
Read one end user's in-app notification feed (GET /notifications): the entries created by sends with `feed` or `showInFeed`. This is what an in-app inbox would show, not delivery status (use get_send_status for that). Paginate by passing `nextCursor` back as `cursor`.
afterlimitbeforecursortemplateIdexternalUserIdrequiredchannelConfigIdGet click stats
Email link click counts grouped by destination URL (GET /clicks) for a `sendId`, a `campaignId`, or both (at least one is required). Only links from email sends made with `trackClicks: true` are counted.
sendIdcampaignIdList webhooks
List the workspace's outbound event webhooks (GET /tenant-webhooks): url, subscribed events, enabled flag and delivery health (consecutiveFailures, lastError). Signing secrets are never returned.
includeArchivedGet webhook
Get one outbound webhook (GET /tenant-webhooks/{id}) including its delivery health (consecutiveFailures, lastDeliveryAt, lastFailureAt, lastError).
webhookIdrequiredCreate webhook
Create an outbound webhook that Ray POSTs signed event payloads to (POST /tenant-webhooks; write scope; Pro plan or higher, otherwise 403). The URL must be public HTTPS. The response contains the signing `secret` exactly once: show it to the user to store (e.g. as an env var), since it can't be read again, only rotated with update_webhook. Read the `webhooks` docs page for the signature scheme.
urlrequirednamerequiredeventsrequiredUpdate webhook
Change a webhook (PATCH /tenant-webhooks/{id}; write scope). Only the fields you pass change; pass at least one. `rotateSecret: true` issues a new signing secret, returned once in the response, and the old secret stops verifying immediately, so the receiving service must be updated at the same time.
urlnameeventsenabledwebhookIdrequiredrotateSecretDelete webhook
Delete (archive) a webhook so it receives no more events (DELETE /tenant-webhooks/{id}; write scope). It cannot be restored through the API; confirm with the user first.
webhookIdrequiredRead Ray docs
Read Ray's documentation as markdown. With no `slug` it returns the index (llms.txt) listing every page; then pass a page slug, i.e. the path after /docs/ without `.md`, such as `sending`, `templates`, `idempotency`, `errors`, `rate-limits`, `webhooks`, `status-and-feeds` or `channels/telegram`. Use it for anything the tool descriptions don't cover.
slug100% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
claude mcp add --transport http ray https://ray-api.gege.mn/mcp --header "Authorization: Bearer $RAY_API_KEY"claude mcp add ray --env RAY_API_KEY=ck_live_... -- npx -y @gege-mn/ray-mcp{ "mcpServers": { "ray": { "command": "npx", "args": ["-y", "@gege-mn/ray-mcp"], "env": { "RAY_API_KEY": "ck_live_..." } } }}{ "mcpServers": { "ray": { "url": "https://ray-api.gege.mn/mcp", "headers": { "Authorization": "Bearer ck_live_..." } } }}{ "mcpServers": { "ray": { "command": "npx", "args": ["-y", "@gege-mn/ray-mcp"], "env": { "RAY_API_KEY": "ck_live_..." } } }}{ "inputs": [ { "id": "ray-api-key", "type": "promptString", "description": "Ray API key (ck_live_...)", "password": true } ], "servers": { "ray": { "type": "http", "url": "https://ray-api.gege.mn/mcp", "headers": { "Authorization": "Bearer ${input:ray-api-key}" } } }}{ "mcpServers": { "ray": { "serverUrl": "https://ray-api.gege.mn/mcp", "headers": { "Authorization": "Bearer ${env:RAY_API_KEY}" } } }}[mcp_servers.ray]command = "npx"args = ["-y", "@gege-mn/ray-mcp"]env = { RAY_API_KEY = "ck_live_..." }[mcp_servers.ray]url = "https://ray-api.gege.mn/mcp"bearer_token_env_var = "RAY_API_KEY"{ "context_servers": { "ray": { "url": "https://ray-api.gege.mn/mcp", "headers": { "Authorization": "Bearer ck_live_..." } } }}{ "mcpServers": { "ray": { "command": "npx", "args": ["-y", "@gege-mn/ray-mcp"], "env": { "RAY_API_KEY": "ck_live_..." } } }}pnpm installpnpm typecheckpnpm test # vitest: bridge end to end against an in-process fake Ray endpointpnpm build # tsup -> dist/index.jspnpm smoke # the built binary over real stdioRAY_MCP_URL=http://localhost:8787/mcp RAY_API_KEY=ck_live_... node dist/index.js