Run Your Outbound from Any AI (MCP)

Manage LinkedIn outbound

• “Which of my campaigns has the highest reply rate over the last 30 days?”
• “Add Mobile Gaming Apps and Esports to the ‘Gaming’ campaign’s industries and re-source.”
• “Show the leads waiting for review on the Publishers campaign, then approve the strong fits.”
• “Loosen the ‘SaaS Founders’ targeting — drop strict keywords and add VP + Director seniority.”
• “Pause the campaign with the lowest accept rate.”

Query your CRM

• “Which open deals haven't had a touchpoint in two weeks?”
• “Summarize my last three meetings with Acme and draft next steps.”
• “What objections did prospects raise about pricing this month?”

Read tools (any key): list_campaigns, get_campaign, get_campaign_analytics, list_prescreen_leads, verify_linkedin_profiles, search_contacts, get_contact, list_deals, list_tasks, get_revenue_summary, list_meetings, get_meeting_summary, and search_call_transcripts (semantic search across every recorded call; Pro plan).

Write tools (keys with the write scope): edit_campaign_targeting, set_campaign_status, create_campaign, add_leads_to_campaign, review_prescreen_leads, import_call_transcript, and create_task.

Every tool, with its exact arguments, is on the tool reference page — point your AI at it and it can learn the whole toolset itself.

The write tools are designed to chain into a full manage-a-campaign loop your AI can run end to end:

1. create_campaign — describe an audience in plain English; Salescadia builds a reviewable draft (or launches it with launch=true).
2. get_campaign → edit_campaign_targeting — read the current filters, then add or remove titles, industries, locations, seniorities, company sizes; set the boolean keyword string, AND/OR keyword-groups, or strict-keyword mode; add exclusions. Saving re-sources with the new filters.
3. list_prescreen_leads → review_prescreen_leads — see who was sourced and approve the real fits (approve enrolls + starts outreach; reject adds to do-not-contact).
4. get_campaign_analytics — read the accept/reply/meeting rates and the trend, then adjust targeting or messaging and go again.

Guardrails hold throughout: sourced leads land in the pre-screen queue when it's on, and no message sends outside the campaign's approval settings — the AI drives the setup and analysis, you keep the send.

1. In Salescadia, open Settings → Integrations and find the Zapier & API Keys card.
2. Click Create API Key. The read scope covers all querying and analytics; add the write scope to let the AI edit targeting, change campaign status, review leads, create campaigns, and import transcripts.
3. Copy the key (it starts with mm_live_) — it is shown once. Treat it like a password; revoke it any time from the same card and every connected AI loses access immediately.

The server lives at https://www.salescadia.com/api/mcp (Streamable HTTP). Two ways to authenticate: sign in with your Salescadia account (OAuth — used by claude.ai) or an API key in the Authorization header (Claude Code, Cursor, custom agents). OAuth connectors are read-only; use an API key with the write scope for the campaign-management tools.

Claude.ai & Claude Desktop

In claude.ai, open Settings → Connectors → Add custom connector, paste https://www.salescadia.com/api/mcp, and click Connect. You'll be sent to Salescadia to sign in and approve access — no key to copy. (This path is read-only; connect an API key below for write tools.)

Claude Code

claude mcp add --transport http salescadia https://www.salescadia.com/api/mcp \
  --header "Authorization: Bearer mm_live_YOUR_KEY"

Cursor / Windsurf / VS Code (mcp.json)

{
  "mcpServers": {
    "salescadia": {
      "url": "https://www.salescadia.com/api/mcp",
      "headers": { "Authorization": "Bearer mm_live_YOUR_KEY" }
    }
  }
}

Custom agents (Claude API, OpenAI, LangChain, etc.)

Any MCP SDK can connect over Streamable HTTP with the same URL + header. The server is stateless — no session setup beyond the standard initialize handshake.

Quick smoke test from a terminal:

curl -s https://www.salescadia.com/api/mcp \
  -H "Authorization: Bearer mm_live_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

You should get back the tool list as JSON. Then ask your assistant something like “list my Salescadia campaigns and show the one with the best reply rate” and watch it call the tools.

• Read-only by default. With a read-scope key, every tool is a read — the AI can list campaigns, inspect targeting, and pull analytics, but change nothing.
• Write tools are scoped and bounded.They only appear on keys you explicitly grant the write scope. They can create campaigns, edit a campaign's targeting and status, review pre-screened leads, and import transcripts — but they never delete your data, and no outreach is ever sent outside a campaign's pre-screen and approval settings. Editing targeting changes who future sourcing finds; it doesn't message anyone. New campaigns are drafts unless you ask to launch.
• Workspace-scoped.A key can only see and act on its own workspace's data — the same isolation our Zapier and Clay integrations use.
• Revocable. Deactivate a key in Settings and every connected AI loses access immediately.
• Anything the AI reads becomes part of that AI conversation — connect assistants you trust with your workspace.

• 401 Unauthorized — the header must be exactly Authorization: Bearer mm_live_...; check the key wasn't revoked or expired.
• A write tool isn't showing up — the key needs the writescope. Create a read+write key in Settings → Integrations and reconnect.
• Transcript search says Pro required — semantic call-transcript search follows the same plan gate as the in-app Ask Anything feature.
• Stuck? Contact support.