Close CRM MCP: Setup, Features & Alternatives

Sales reps should not need six filters and a CSV export to answer a basic pipeline question. Close CRM MCP gives an AI Sales Assistant a direct route into Close, where it can retrieve sales records, interpret activity data, and complete authorized CRM actions from a conversational prompt.

The official server uses Model Context Protocol to expose live leads, contacts, opportunities, emails, calls, tasks, notes, and meeting transcripts to compatible AI clients. It extends Close beyond its native workspace, but its outbound-heavy setup will not fit every revenue team. The real test lies in the actions it supports, the control it provides, and how it compares with folk CRM MCP.

What Is Close CRM MCP?

Close CRM MCP is a hosted server that translates requests from an AI client into structured actions inside Close. It uses Model Context Protocol, an open standard that allows AI models to interact with external software and live business data.

The connection runs through OAuth rather than a manually shared API key. During authorization, access can be limited to three permission levels:

• Read Close data
• Write or modify Close data
• Delete Close data

Close MCP is separate from Chloe, the AI sales agent built into the CRM. Chloe handles workflows such as calling leads, qualifying prospects, preparing follow-ups, and updating records. The MCP server instead allows external assistants such as Claude or ChatGPT to work with Close data and tools from their own interfaces.

What Can Close MCP Do? 5 Main Use Cases

Close MCP supports both data retrieval and CRM actions. Its toolset covers the records and activities that outbound sales teams handle throughout the day.

1. Search Leads, Contacts, and Opportunities

→ AI clients can retrieve leads, contacts, organizations, and opportunities using natural-language criteria. Requests can combine Close fields such as status, owner, value, confidence, creation date, or recent activity.

A sales manager could ask for every open opportunity above a specific value with no activity during the past 14 days, then group the results by owner.

2. Create and Update CRM Records

→ Close MCP can create leads and modify existing records when write access is enabled. Supported actions include updating lead details, editing opportunity fields, changing confidence levels, and adding information found during external research.

This makes the connection useful for data cleanup, record enrichment, and bulk updates that would otherwise require manual editing or a custom script.

3. Analyze Sales Activity and Pipeline Data

→ The server can retrieve calls, emails, meetings, opportunities, and other logged activities for analysis. An AI assistant can calculate activity breakdowns, summarize pipeline movement, compare rep performance, or identify deals with weak engagement.

The analysis remains grounded in current Close data instead of an exported spreadsheet that becomes outdated after download.

4. Manage Tasks, Notes, and Follow-Ups

→ Authorized AI clients can create tasks, add notes, and organize next steps around existing leads. Sales teams can turn a pipeline review into assigned follow-up actions without switching back to the CRM after every decision.

5. Retrieve Meeting Transcripts

→ Close MCP can fetch transcripts captured by Close Notetaker. An AI assistant can summarize the conversation, extract objections, identify commitments, and convert agreed next steps into CRM notes or tasks.

Transcript access creates a direct path from a sales conversation to structured follow-up work, with no manual copy and paste.

How Does Close CRM MCP Work?

Close MCP acts as an intermediary between the AI client and the Close API. The AI model interprets a request, selects the appropriate MCP tool, and sends a structured call to the hosted server. Close then returns the requested data or performs the authorized action.

The workflow follows five steps:

1. The user connects Close to an MCP-compatible AI client.
2. OAuth verifies the Close account and selected organization.
3. The user grants read, write, and optional delete permissions.
4. The AI client selects a Close MCP tool based on the prompt.
5. Close executes the request and returns the result to the conversation.

Permissions determine what the AI can do. A read-only connection can search records and analyze sales activity without changing CRM data. Write access enables record creation and updates, while delete access permits destructive actions and should remain limited to workflows that genuinely require it.

The hosted architecture removes the need to deploy or maintain a local server. Access can also be revoked from the Authorized Apps settings inside Close.

How to Set Up Close CRM MCP?

Close hosts its remote MCP server, so most connections only require OAuth authorization. Custom clients can also authenticate with a Close API key.

Close Account Requirements

Before connecting the server, the setup requires:

• An active Close account
• Access to the relevant Close organization
• A compatible MCP client
• Permission to authorize third-party applications
• An API key for clients that do not support Close OAuth

The official remote endpoint supports Streamable HTTP and OAuth 2.0 Dynamic Client Registration.

Connect Close to Claude

Close appears as an official connector across the Claude ecosystem.

1. Open the Connectors menu in Claude.
2. Search for Close.
3. Select Connect.
4. Sign in to Close.
5. Choose the correct organization.
6. Approve the required permission scope.
7. Enable the Close tools in the conversation.

Claude’s connector library requires a paid plan on the web app. The Close connector can also be added through the Claude desktop app on free plans.

Connect Close to ChatGPT

Close provides an official app for ChatGPT. The previous Close CRM GPT is being retired in favor of this newer connection.

1. Open the official Close app in ChatGPT.
2. Start a conversation with the app.
3. Sign in to the Close account when prompted.
4. Select the appropriate organization.
5. Review and approve the requested access.
6. Test the connection with a read-only request.

Configure Read, Write, and Delete Permissions

Close separates MCP access into three scopes:

• mcp.read: Read CRM data without making changes
• mcp.write_safe: Read, create, and update CRM data
• mcp.write_destructive: Read, write, and delete CRM data

Read-only access provides the safest starting point for reporting and analysis. Write access can be added once the workflow has been tested. Delete permissions should remain disabled unless a defined process requires them.

Close CRM MCP Pricing

Close does not charge a separate fee for access to its MCP server.

Close CRM MCP: Strengths and Limitations

4 Best Close CRM MCP Alternatives in 2026

👉folk 無料で試す

1.folk

評価

⭐⭐⭐⭐⭐ (G2)

概要

folk is a lightweight CRM built around relationships rather than call volume. Its hosted MCP server allows AI assistants to search people and companies, review interactions, create records, update deals, add notes, and work across structured CRM groups.

Where Close concentrates on dialers and outbound execution, folk brings together contact enrichment, relationship history, personalized messaging, and flexible pipelines. An AI assistant can retrieve the context behind a relationship before drafting a follow-up or recommending the next action.

The simpler interface and predictable pricing make folk particularly relevant for startups, agencies, recruiters, investors, and partnership teams. It provides native MCP access without pushing every workflow into a conventional sales funnel.

Best Features

• Native hosted MCP server
• Secure OAuth authentication
• Read and write access to CRM records
• People and company enrichment
• Relationship and interaction timelines
• Flexible groups and pipelines
• AI-powered personalized messages
• Email sequences with reply detection
• Meeting summaries
• Next Best Action recommendations
• Gmail, Outlook, LinkedIn, and WhatsApp integrations

価格設定

• 標準プラン：1ユーザーあたり月額24ドル
• Premium：1ユーザーあたり月額48ドル
• Custom: From $80/user/month
• 無料トライアル：14日間

👉folk 無料で試す

✅ folk is the strongest Close MCP alternative for teams that value relationship context, prospect research, and personalized outreach more than built-in calling infrastructure.

2 アティオ

評価

⭐⭐⭐⭐ (G2)

概要

Attio is a customizable CRM for startups and revenue teams that need more control over their data model. Its native MCP server allows compatible AI clients to search, query, create, and update workspace records.

Unlike Close, Attio does not center the product around built-in calling and SMS. It focuses on connected customer data, custom objects, flexible attributes, workflows, and collaboration. This makes it better suited to teams building a tailored revenue system rather than running a call-heavy outbound operation.

Attio provides more structural flexibility than Close, but that freedom brings extra setup work. Teams need to define objects, views, attributes, and automations before the CRM fully matches their process.

Best Features

• Native hosted MCP server
• Search, query, and write actions
• Custom objects and attributes
• Flexible workspace views
• Email and calendar synchronization
• 自動化されたワークフロー
• Contact and company enrichment
• Granular permissions
• API and developer platform
• Integrations with ChatGPT, Claude, Notion, Slack, and Clay

価格設定

• 無料：最大3ユーザーまで0ドル
• Plus: $35/user/month
• プロ：79ドル/ユーザー/月
• エンタープライズ: カスタム価格設定

✅ Attio fits revenue teams that prefer a highly configurable CRM data model over Close’s built-in communication stack.

3.Pipedrive

評価

⭐⭐⭐⭐ (G2)

概要

Pipedrive is a sales CRM built around visual pipeline management. Its native MCP server connects ChatGPT, Claude, and other compatible AI assistants to live deals, contacts, organizations, leads, activities, and notes.

AI clients can search records, create and update CRM data, analyze pipelines, and automate supported tasks through natural-language prompts. The server follows existing Pipedrive roles and visibility settings, while recorded actions remain available in the change log.

Compared with Close, Pipedrive offers a more visual and conventional deal-management experience. Close remains stronger for teams that depend on native calling, SMS, and high-volume dialers, while Pipedrive suits smaller sales teams that prioritize pipeline visibility and ease of adoption.

Best Features

• Official native MCP server
• ChatGPT and Claude compatibility
• Secure OAuth authentication
• Live deal, contact, lead, and organization data
• CRM search and write actions
• Existing permission inheritance
• Action audit trail
• 視覚化された販売パイプライン
• Activity and follow-up management
• ワークフロー自動化

価格設定

• Lite: From $14/user/month
• Growth: From $39/user/month
• Premium: From $59/user/month
• Ultimate: From $79/user/month
• 無料トライアル：14日間

✅ Pipedrive is a practical Close MCP alternative for teams that want straightforward pipeline management without paying for a call-centric CRM stack.

4. Twenty

評価

⭐⭐⭐⭐

概要

Twenty is an open-source CRM that includes a native MCP server with every cloud workspace. ChatGPT, Claude, Cursor, and other compatible clients can connect through OAuth to read and write CRM data.

The platform includes people, companies, opportunities, notes, tasks, workflows, and custom objects. Technical teams can use the managed cloud product or deploy the open-source version on their own infrastructure.

Twenty costs considerably less than Close and provides more control over the underlying software. However, it lacks Close’s mature calling, SMS, email, and sales engagement infrastructure. Self-hosting also transfers maintenance, security, and upgrade responsibilities to the internal team.

Best Features

• Native MCP server on every cloud plan
• Natural-language read and write actions
• OAuth authentication
• Open-source codebase
• Custom objects and fields
• Flexible record views
• ワークフロー自動化
• API and webhooks
• Cloud deployment
• Self-hosting option

価格設定

• Cloud Pro: $9/user/month
• Organization: $19/user/month
• Self-hosted: Open-source deployment

✅ Twenty works best for technical teams that want low-cost MCP access, CRM customization, and infrastructure control without Close’s outbound communication suite.

Which Close MCP Alternative Should Sales Teams Choose?

The right choice depends on how the team sells and how much control it needs over CRM structure:

• Choose folk CRM for relationship-driven sales, recruiting, partnerships, fundraising, and personalized outreach. It combines live CRM context, enrichment, flexible pipelines, and native MCP access in a simpler workspace.
• Choose Attio for custom objects, flexible attributes, and a CRM architecture built around a specific revenue process.
• Choose Pipedrive for visual pipeline management, straightforward deal tracking, and familiar sales workflows.
• Choose Twenty for an affordable open-source CRM with native MCP access and optional self-hosting.

For most startups and lean teams, folk offers the strongest balance of relationship intelligence, prospect research, personalized outreach, and ease of use. Close remains the better option when built-in calling, SMS, and dialers sit at the center of daily sales activity.

Close MCP vs folk CRM MCP

Close and folk both provide hosted MCP servers with read and write access to live CRM data. Close centers its product on outbound activity and built-in communication, while folk places more emphasis on relationship context, enrichment, and personalized outreach.

結論

Close CRM MCP brings live pipeline data and CRM actions into ChatGPT, Claude, and other compatible AI clients. Its official hosted server, granular permission scopes, meeting transcript access, and native sales communication stack make it a solid choice for outbound teams.

The product becomes less compelling when relationship context matters more than call volume. folk CRM combines native MCP access with enrichment, interaction history, flexible pipelines, and personalized outreach, making it a more balanced option for startups and lean teams.

よくあるご質問

Does Close CRM Have an Official MCP Server?

Yes. Close provides an official hosted MCP server that connects compatible AI assistants to live CRM data. The remote server supports OAuth 2.0 and Streamable HTTP.

Which AI Assistants Work With Close MCP?

Close MCP works with ChatGPT, Claude, Claude Code, Cursor, VS Code, n8n, and other clients that support remote MCP connections. Available permissions and setup methods vary by client.

Can Close MCP Create and Update CRM Records?

Yes. The mcp.write_safe scope allows an AI client to read, create, and update Close data. The separate mcp.write_destructive scope also enables deletion.

Is Close MCP Included With Every Plan?

Close does not list MCP as a separate paid add-on. Close CRM plans start at $9 per user per month with annual billing, while subscriptions to ChatGPT, Claude, or other external AI clients remain separate.

What Is the Best Alternative to Close CRM MCP?

folk CRM is the strongest alternative for startups and lean teams that prioritize relationship intelligence, enrichment, flexible pipelines, and personalized outreach. Attio offers deeper data-model customization, Pipedrive focuses on visual deal management, and Twenty provides an affordable open-source option.