Back to docs
Kilo integrates with Claude through the Model Context Protocol (MCP). Once connected, your sales team can ask Claude natural language questions about your leads and qualification data — no dashboards or queries needed.
Once Kilo is connected, anyone on your team can ask questions like:
"Show me leads where a champion is identified"
"What pain did the Stripe visitor describe?"
"Which leads have metrics captured but no economic buyer?"
"Give me the full MEDDPICC scorecard for the Notion lead"
"List hot leads from the last 7 days"
"What dimensions are we consistently missing?"
Claude uses Kilo's MCP tools to pull live data from your qualification database and answer in natural language.
Step 1 — Generate an API key
Go to Account → Integrations in your Kilo dashboard. Click "Generate API key." Copy the key immediately — it's only shown once.
Your API key starts with
Step 2 — Copy the server URL
The MCP server URL is shown on the same page. It looks like:
Step 3 — Add to Claude
Open Claude (claude.ai or the desktop app)
Go to Settings → Integrations
Click Add custom integration
Paste the Server URL from step 2
Set authentication to Bearer token
Paste your API key from step 1
Click Save
That's it. Your Kilo tools will appear in Claude conversations immediately.
Kilo exposes 6 tools to Claude:
get_framework_schema — Returns your configured qualification framework (MEDDPICC, BANT, or SPICED) and all dimension keys. Claude calls this first to understand your setup.
get_recent_leads — Returns scored lead briefs from the last N days, sorted by score. Use this for questions like "Show me this week's leads" or "What hot leads came in?"
filter_leads_by_dimension — The core query tool. Filters leads by whether a specific dimension is identified or missing. Handles questions like "List leads where champion is identified" or "Show leads where Metrics is missing."
filter_leads_by_dimension_combo — Multi-dimension gap analysis. Handles complex queries like "Which leads have a champion but no economic buyer?" or "Leads with pain and metrics but no decision process."
get_dimension_value — Returns the extracted content and verbatim conversation excerpt for a specific dimension on a specific lead. Answers questions like "What pain did the Stripe visitor describe?" or "Who is the champion on the Notion deal?"
get_lead_scorecard — Full qualification scorecard for one lead — all dimensions with status, value, confidence, and conversation excerpts.
Your API key is scoped to your organization — you can only query your own data
Keys are hashed with SHA-256 before storage — we never store the plain key
You can revoke keys anytime from Account → Integrations
Each key has a "last used" timestamp so you can see if it's active
Start with the framework schema. If Claude doesn't know your framework, it can't query dimensions correctly. Ask "What framework is configured?" first.
Be specific about dimensions. Instead of "show me qualified leads," try "show me leads where Identified Pain and Champion are both identified." Claude maps these to the exact dimension keys.
Use lead names. If you know the company name, say "Show me the scorecard for the Notion lead" instead of using session IDs.
Does this work with Claude.ai and Claude desktop? Yes, both. Any Claude interface that supports custom integrations.
Can multiple team members use the same API key? Yes, but we recommend one key per person so you can track usage and revoke individually.
Is my data sent to Anthropic? Claude sends tool requests to Kilo's MCP server, which returns structured data. The data passes through Claude for the response — the same as any Claude integration. No data is stored by Anthropic beyond their standard data handling policies.
What if I use BANT instead of MEDDPICC? The integration works with any framework configured in Kilo. Claude automatically adapts to your dimension keys.
Claude Integration
Connect Kilo to Claude
Kilo integrates with Claude through the Model Context Protocol (MCP). Once connected, your sales team can ask Claude natural language questions about your leads and qualification data — no dashboards or queries needed.
What you can ask Claude
Once Kilo is connected, anyone on your team can ask questions like:
Claude uses Kilo's MCP tools to pull live data from your qualification database and answer in natural language.
Setup (5 minutes)
Step 1 — Generate an API key
Go to Account → Integrations in your Kilo dashboard. Click "Generate API key." Copy the key immediately — it's only shown once.
Your API key starts with
kilo_sk_ and looks like: kilo_sk_a3f8b9c2d1e4...Step 2 — Copy the server URL
The MCP server URL is shown on the same page. It looks like:
https://mcp.kilo-sales.com/mcpStep 3 — Add to Claude
That's it. Your Kilo tools will appear in Claude conversations immediately.
Available tools
Kilo exposes 6 tools to Claude:
get_framework_schema — Returns your configured qualification framework (MEDDPICC, BANT, or SPICED) and all dimension keys. Claude calls this first to understand your setup.
get_recent_leads — Returns scored lead briefs from the last N days, sorted by score. Use this for questions like "Show me this week's leads" or "What hot leads came in?"
filter_leads_by_dimension — The core query tool. Filters leads by whether a specific dimension is identified or missing. Handles questions like "List leads where champion is identified" or "Show leads where Metrics is missing."
filter_leads_by_dimension_combo — Multi-dimension gap analysis. Handles complex queries like "Which leads have a champion but no economic buyer?" or "Leads with pain and metrics but no decision process."
get_dimension_value — Returns the extracted content and verbatim conversation excerpt for a specific dimension on a specific lead. Answers questions like "What pain did the Stripe visitor describe?" or "Who is the champion on the Notion deal?"
get_lead_scorecard — Full qualification scorecard for one lead — all dimensions with status, value, confidence, and conversation excerpts.
Security
Tips
Start with the framework schema. If Claude doesn't know your framework, it can't query dimensions correctly. Ask "What framework is configured?" first.
Be specific about dimensions. Instead of "show me qualified leads," try "show me leads where Identified Pain and Champion are both identified." Claude maps these to the exact dimension keys.
Use lead names. If you know the company name, say "Show me the scorecard for the Notion lead" instead of using session IDs.
Common questions
Does this work with Claude.ai and Claude desktop? Yes, both. Any Claude interface that supports custom integrations.
Can multiple team members use the same API key? Yes, but we recommend one key per person so you can track usage and revoke individually.
Is my data sent to Anthropic? Claude sends tool requests to Kilo's MCP server, which returns structured data. The data passes through Claude for the response — the same as any Claude integration. No data is stored by Anthropic beyond their standard data handling policies.
What if I use BANT instead of MEDDPICC? The integration works with any framework configured in Kilo. Claude automatically adapts to your dimension keys.