Docs

Connect once. Keep the operating path in one place.

Configure Claude or Gemini CLI, verify the live endpoint, inspect account state, and route security or support questions without crossing into the self-host documentation lane.

Start free — no card Open account

Primary endpoint

The default commercial path is one hosted endpoint plus one API key.

https://mcp.revasserlabs.com/mcp
View plans Open account

Status surfaces

  • /status is the browser-safe status contract for version, tool counts, token modes, and support links.
  • /health is liveness only.
  • /api is the machine-readable service directory payload.

Compatibility quick matrix

  • Claude Desktop / Claude Code: strongest fit for self-host through stdio with npx -y @jtalk22/slack-mcp --setup.
  • Gemini CLI: supported on the hosted HTTP path when the buyer wants remote MCP outside Claude.
  • Hosted rollout: use deployment review when the real question is runtime ownership, support envelope, or buyer review.

Claude config

{ "mcpServers": { "slack-mcp-cloud": { "url": "https://mcp.revasserlabs.com/mcp", "headers": { "Authorization": "Bearer stmh_..." } } } }

Gemini CLI config

~/.gemini/settings.json { "mcpServers": { "slack-mcp-cloud": { "httpUrl": "https://mcp.revasserlabs.com/mcp", "headers": { "Authorization": "Bearer stmh_..." } } } } CLI gemini mcp add --transport http --header "Authorization: Bearer stmh_..." slack-mcp-cloud https://mcp.revasserlabs.com/mcp

Cloud vs self-host

  • Self-host: 21 tools, npm or Docker, full control over runtime and tokens. Contract-generating inference stays hosted.
  • Free: $0 (no card) — 2000 requests + 25 AI tool calls per month.
  • Pro: $19/mo for unlimited requests, unlimited inference tools, opt-in encrypted credential storage, scheduled delivery, and signed webhooks.
  • Team: $49/mo flat for 5 workspaces + shared profiles.
  • Safeguard: $199/mo (waitlist) — agent approval gates, searchable workspace memory, and an audit log (in development).

Workflow operations

  • Save: call slack_workflow_save with a name, workflow kind, channel set, timezone, and optional cadence.
  • List: call slack_workflows to retrieve the current tenant-scoped profiles.
  • Run: call slack_workflow_brief with profile_name; direct workflow-kind calls remain supported.
  • Schedule: paid active tenants with persistent Slack credentials can deliver daily at 8am or Monday at 8am in the profile timezone.
  • Deliver: Pro and Team can add a public HTTPS incident/status webhook. Payloads use a timestamped P-256 signature verifiable with the returned public JWK.

Deployment modes

  • Local stdio: lowest-ops path for one operator using Claude Desktop or Claude Code.
  • Local web: browser-first local mode when MCP transport is not the main entry point.
  • Hosted HTTP: managed remote path when runtime, support, and buyer review matter more than infrastructure ownership.
  • Worker / registry distribution: reserve for buyers that explicitly need registry-hosted transport and release validation.

Support and buyer review

  • Use deployment review for rollout, plan fit, and security questions.
  • Use security for buyer-facing controls, storage, and analytics review.
  • Use Ops engagement for SLA, custom retention, SOC2 evidence, multi-tenant isolation.
  • Use official vs managed when the team is deciding between Slack's official/self-host path and the managed Cloud path.
  • Use workflows for the 5 workflow profile types and how they pair with the AI tools.
  • Use support for hosted operations, billing, and follow-up.
  • Use privacy for token modes, retention, and analytics disclosure.

Operating note

Operated by Revasser. Public GitHub docs and issues stay focused on self-hosted usage; managed rollout and Cloud support stay on these hosted routes.