Supervised• Every action logged
Connect ChatGPT or Claude to Your CRM Pipeline — via MCP
HelloGrowthCRM ships a public Model Context Protocol (MCP) server. Point any MCP-compatible AI client at your CRM and query leads, update deals, trigger sequences, and pull analytics — in plain language. The biggest defensibility moat in CRM, now available at SMB pricing.
Claude ↔ HelloGrowthCRM MCP
# User prompt to Claude
> "What deals in my pipeline have been stalled for >14 days?"
# Claude calls MCP tool
tool: get_pipeline
filter: { days_since_update: ">14", stage: "not closed" }
# CRM responds
→ 7 deals returned: Acme Corp ($45K, 22 days), ...
# Follow-up action
> "Create a follow-up task for each of those deals due Monday"
tool: create_task × 7
→ 7 tasks created in CRM. Logged as MCP action by Claude.
Available MCP Tools
10 CRM operations available as MCP tools today. More added every sprint.
get_pipeline
Fetch your full deal pipeline with filters
update_deal
Update deal stage, value, or custom fields
create_contact
Add a new CRM contact with enrichment
get_lead_score
Retrieve AI lead score for any contact
trigger_sequence
Start an outreach sequence for a contact
log_activity
Log a call, meeting, or note to a record
get_analytics
Fetch revenue and pipeline analytics
search_contacts
Natural-language contact search
create_task
Create a CRM task for a rep or team
get_meeting_notes
Retrieve AI-generated meeting summary
How to Connect
1
Install the MCP server
Add the HelloGrowthCRM MCP server URL to your AI client config. Available as a hosted endpoint or self-hosted Docker image.
2
Authenticate with your API key
Use your HelloGrowthCRM API key. Scopes are configurable — give the AI client read-only or read-write access per your comfort level.
3
Your AI client can now see your CRM
Ask ChatGPT 'What deals are stalled?' or tell Claude 'Update the Acme deal to Proposal stage' — it queries your live CRM data.
4
Every action is logged
Every MCP-initiated action is logged in the CRM audit trail with the AI client name and timestamp. Full governance, no black box.
Try it — Connect, Ask, Trace
See how a connection looks for your assistant of choice, what users actually ask in plain English, and the full tool-call trace your reviewers see in the audit log.
Pick the assistant you use most. Add the snippet to its config and it gains read/write access to your HelloGrowthCRM workspace through the MCP server.
~/Library/Application Support/Claude/claude_desktop_config.json
{
"mcpServers": {
"hellogrowthcrm": {
"command": "npx",
"args": ["-y", "@hellogrowthcrm/mcp-server"],
"env": {
"HGCRM_API_KEY": "<your-api-key>",
"HGCRM_WORKSPACE": "<your-workspace-id>"
}
}
}
}Compatible AI Clients
ChatGPT (via GPT Actions)
Claude (Anthropic MCP client)
Gemini Advanced
Cursor IDE
Continue.dev
Zed editor
Any OpenAI-compatible client
Custom Python/Node agents
Safety & Governance
Every MCP-initiated action is logged with AI client identity and timestamp
Configurable scope: read-only API key, or read+write with action limits
PII fields can be masked from AI client responses
Rate limiting per API key prevents runaway automation
One-click revoke API access from your admin panel
The First CRM with a Public MCP Server
Get your MCP API key in under 5 minutes. Free with every HelloGrowthCRM account.