Dashboard

Peerbound MCP: Quick Start Guide

Last updated 5 days ago

Please reach out to your CSM to have the MCP enabled for your organization.

Peerbound is a customer proof platform that catalogs positive experiences from your customers — stories, G2 reviews, and moments.

Model Context Protocol (MCP) is simply a way to plug a tool like Peerbound into an AI assistant, so it can directly access your data or take actions on your behalf.

The Peerbound MCP connects your Peerbound customer proof library directly to your AI assistant, giving your assistant on-demand access to quotes, case studies, and reviews, allowing it build proposals, emails, and pitches with verified customer data.

For a full reference of available MCP tools and parameters, see Peerbound MCP Tools.

When to use the MCP

Use the MCP when you're drafting a proposal, email, or pitch and need relevant customer proof surfaced directly in your workflow — no switching between applications. It's especially useful for complex searches where you need to filter by industry, use case, or company type simultaneously, and for retrieving detailed, structured content like full case study summaries, quotes, and reviews in a single response.

Example prompts

  1. "What have our customers been saying about us? Analyze recent calls and pull out 5 key themes."

  2. "Find approved customer proof points for [industry] customers using [feature]."

  3. "Create a battlecard for us vs. [Competitor] company using customer proof."

  4. "I'm prospecting into [company]. Have we served similar companies? Share approved quotes only."

  5. "Help me prep for the [Prospect] deal."

Connecting Peerbound to AI tools

Depending on the AI tool you're using, you may need to periodically reauthenticate the Peerbound MCP to maintain the connection.

Connecting with Claude

Option 1: Claude Admin configuring Peerbound for the entire organization

Choose this option if you want to make Peerbound available to your sales and marketing teams in Claude. You'll need a Claude Owner or Primary Owner to complete the setup.

Step 1 – Add Peerbound as an organization connector

  1. Install Peerbound from the Connector Directory.

Note: On Team and Enterprise plans, only Owners or Primary Owners can add connectors. Admin access alone is not sufficient.

Step 2 – Each user connects their individual account

Once Peerbound has been added org-wide, every user who wants access must authenticate individually:

  1. Select Customize in the left side bar

  2. Select Connectors from the menu

  3. Scroll to the bottom to find Peerbound and Connect

  4. You will be prompted to authenticate your individual account

Note: Adding Peerbound at the org level does not automatically grant access to anyone. Each person must complete this step before they can use it.

Option 2: Configuring Peerbound for your individual account

Choose this option if you want to get set up quickly without going through your Claude admin. Keep in mind that adding connectors may require certain permissions depending on your organization's Claude plan – check with your admin if you're unsure.

  1. Install Peerbound from the Connector Directory.

  2. Follow the prompts to authenticate your account

Note: This connects Peerbound for your account only. It will not be visible or available to anyone else in your organization.

Connecting with ChatGPT

In order to add Peerbound to ChatGPT you must add Peerbound as a "custom connector".

ChatGPT supports custom connectors in "Developer Mode", which is available in beta to Pro, Plus, Business, Enterprise, and Education accounts on the web. To connect:

  1. Enable Developer Mode in Settings → Apps → Advanced settings.

  2. Go to ChatGPT Apps settings and click "Create app" to add the Peerbound remote MCP server URL: https://mcp.peerbound.com/api/mcp

  3. Enter the Base URL and configure authentication.

See OpenAI's documentation for more details.

Connecting to Codex

1. Run the following commands in your terminal.

codex mcp add peerbound --url https://mcp.peerbound.com/api/mcp

2. Next, run 

codex mcp list

You should see peerbound listed among your configured MCP servers.

You can also open Codex and run:

/mcp

to view available MCP servers and tools.

3. Authenticate: Plugins → MCPs → Peerbound Settings → Authenticate

When prompted, complete the authentication flow using your Peerbound credentials. Depending on your organization's configuration, this may use OAuth or an API token.

Connecting with Microsoft Copilot Studio

Copilot Studio supports connecting to existing MCP servers directly within the agent builder. You'll need access to a Copilot Studio agent to complete this setup.

Step 1 - Create a new agent in Microsoft Copilot Studio

  1. Name it "Peerbound Assistant"

  2. Select Add a tool → New tool → Model Context Protocol

Step 2 – Configure the Peerbound MCP server

  1. Fill in the following fields in the MCP onboarding wizard:

    • Server name: Peerbound

    • Server description: Access Peerbound customer proof — quotes, case studies, and reviews — to build proposals, emails, and pitches with verified customer data.

    • Server URL: https://mcp.peerbound.com/api/mcp

  2. Select your authentication method:

    • OAuth 2.0 — recommended for individual users authenticating with their own Peerbound credentials

    • API key — recommended for automated workflows and system accounts (generate your API key from https://app.peerbound.com/settings/api-keys)

  3. Select Create

Step 3 – Add the server to your agent

  1. On the Add tool dialog, select Create a new connection

  2. Select Add to agent to finish

Note: A Peerbound admin must first enable MCP access for your organization. Contact your CSM if it hasn't been enabled yet.

See Microsoft's documentation for more details.

Connecting with Google Gemini (Enterprise)

Gemini Enterprise supports connecting to existing MCP servers directly within the Google Cloud Console. You'll need organizational admin access to Gemini Enterprise (Vertex AI / Discovery Engine) to complete this setup.

Step 1 – Add Peerbound as a custom data store

  1. Go to the Google Cloud Console and navigate to Gemini Enterprise -> Data stores.

  2. Click Create data store.

  3. On the data source configuration page, search for and select Custom MCP Server.

Step 2 – Configure the Peerbound MCP server

  1. Fill in the following fields in the MCP onboarding wizard:

    • Server name: Peerbound

    • Server description: Access Peerbound customer proof — quotes, case studies, and reviews — to build proposals, emails, and pitches with verified customer data.

    • Server URL: (https://mcp.peerbound.com/api/mcp)

  2. Select your authentication method:

    • OAuth 2.0 (Recommended) — Provide your base Authorization URL, Client ID, and Client Secret to allow individual users to authenticate securely with their own Peerbound credentials.

    • API key — Recommended for automated workflows and background processing. Provide the key using a Bearer token format. You can generate your API key from (https://app.peerbound.com/settings/api-keys).

Step 3 – Authorize and enable the tools

  1. Once the server data store is successfully created, select it from your list and navigate to Actions.

  2. Click Reload custom actions to fetch the available tools exposed by Peerbound.

  3. Check the authorization boxes next to the tools you want to make available to Gemini, and click Enable actions to finish.

See Google Cloud's documentation for more details.

Connecting Peerbound to AI workflow tools

The Peerbound MCP supports two authentication methods: API key or custom connector with OAuth.

API Key Authentication — for system accounts

Use this method when connecting Peerbound to an automated workflow, agent pipeline, or integration where no human user is signing in — for example, connecting to tools like n8n, Gumloop, or Relevance AI.

A Peerbound admin must generate the API key from the Company Settings page at https://app.peerbound.com/settings/api-keys.

Include the key in the Authorization header:

Authorization: Bearer <your-api-key>

API keys grant access at the account level, not on behalf of any individual user.

Optional request headers

Alongside Authorization, a connection can send up to three headers that control what Peerbound returns. All are optional; each has a default.

Value

Default

What this controls

When to use this

X-Attribute-Anonymous

true

whether anonymous quotes are returned at all

if you are building a connection that should exclusively serve approved content set this to false

X-Include-Anonymous

true

whether anonymous quotes carry the real company and contact.

if you are building a connection where you want access to anonymous quotes, but no chance of accidentally revealing the customer company or name set this to false

X-Minimum-G2-Stars

4

the lowest star rating of review to return

if you only want positive reviews set this to a higher value if you want all reviews set this to a lower value

Header values

A header counts as enabled only when its value is true, 1, or yes. Any other value, including False or no, is read as disabled. Header names are not case-sensitive.

OAuth Authentication — for individual users

Use this method when a human user is connecting Peerbound through an AI assistant like Claude or ChatGPT (see "Connecting Peerbound to Claude or ChatGPT"). Users authenticate with their own Peerbound credentials.

Access through MCP reflects the same permissions a user has in Peerbound — or, for Sales users who don't have direct app access, their Slack permissions. If a customer story or moment isn't accessible to them in their existing tools, they won't be able to retrieve it via MCP either. This makes OAuth the right default for sales and marketing teams using AI assistants day-to-day.

Related Articles
Peerbound MCP Tools Microsoft Teams Integration Instructions G2 Integration Instructions Slack Integration Instructions