Connect AI agents to your testimonials with an MCP server
A testimonial MCP server lets Claude, ChatGPT, or your own AI agent read, add, and manage testimonials by chat.
You connect once, choose scoped permissions, and then ask in plain language instead of logging in and copy-pasting.
Why testimonials stay out of reach for AI tools
Collecting reviews is only half the work. Once they exist, an AI assistant still cannot use them if they sit behind a login. The usual workarounds are slow: copy-pasting into chat, asking a developer to wire a REST client and store an API key, or opening the dashboard for every small edit such as renaming a wall or fixing a typo. An MCP connection removes that friction by giving the assistant a standard, permission-scoped way to reach the same walls.
Connect once, then ask in plain language
Point Claude or ChatGPT at the MCP address, sign in with the account you already have, and tick only the permissions you want. No API key is needed for those chat clients. You can disconnect at any time from the AI client's settings. The illustrative conversation in protected module code0001 shows a typical request for recent testimonials; exact wording from your assistant will vary.
Two connection paths: chat sign-in or agent key
Chat assistants: open Settings, then Connectors, add a custom connector, paste the MCP address, click Connect, sign in, and choose what it can access. This takes about two minutes. Autonomous agents: point the agent at the same address, generate a dedicated fw_ag_ key from the Connect AI Agents page, and send it as Authorization: Bearer fw_ag_your_key on every request. Examples include n8n workflows, the OpenAI Agents SDK, the Claude Agent SDK, or a custom script built on the MCP SDK. Use this checklist to decide: if the client can click a sign-in button, use OAuth; if it runs unattended, use the dedicated key.
Practical prompts you can paste into chat
Try asking for a summary of recent testimonials, adding a 5-star testimonial with a named author and quote, searching for mentions of support, finding testimonials that help landing-page conversions, listing submissions waiting for review, fixing a typo and updating an author's job title, or renaming a wall. Your assistant should report what it did and confirm the result. Before you ask, decide which permission group the task needs: read-only for summaries and searches, add-and-edit for changes.
Built on an open standard, not a bespoke API
Because the server uses the Model Context Protocol, any MCP-compatible client can connect without custom integration code. Listed clients include Claude, ChatGPT, n8n, the OpenAI Agents SDK, the Claude Agent SDK, and custom MCP SDK agents, plus any other client that supports custom connectors. For machine clients, the reference details are: human auth via OAuth 2.1 sign-in for Claude and ChatGPT with no key; agent auth via Authorization: Bearer <fw_ag_key>; discovery via GET /.well-known/oauth-protected-resource; and on a rate-limit response, retry after about 60s.
Controls that keep the assistant inside the lines
Grant only the abilities you intend to allow, from viewing walls to editing testimonials. Nothing runs outside those permissions. Agents cannot delete a testimonial or a wall; deletion remains a dashboard-only action. To stop access, revoke the connector from your AI client or regenerate the autonomous agent key in one click.
| Plain-English action | Tool | Permission needed |
|---|---|---|
| List your walls | list_walls | View your walls |
| Open one wall, with its testimonial count | get_wall | View your walls |
| List testimonials on a wall | list_testimonials | Read your testimonials |
| Open one testimonial | get_testimonial | Read your testimonials |
| Search testimonials by text, wall, rating, or type | search_testimonials | Read your testimonials |
| Count walls, testimonials and pending submissions | get_account_usage | View account insights |
| See submissions waiting for review | list_collected_submissions | View submissions |
| Add a text testimonial to a wall | add_text_testimonial | Add & edit testimonials |
| Edit a testimonial's wording, author, rating or links | update_testimonial | Add & edit testimonials |
| Rename a wall or change its description | update_wall_settings | Edit your walls |
Permissions are chosen at connection time; nothing runs outside the granted set.
Frequently asked questions
What is a testimonial MCP server?
It connects an AI assistant or agent directly to your testimonial walls using the Model Context Protocol, an open standard for linking AI tools to external data. After connection, the assistant can read, add, and tidy testimonials through chat without opening the dashboard.
Do I need to know how to code to connect Claude or ChatGPT?
No for Claude and ChatGPT. Those clients sign in with your existing account, with no key to generate or paste. Coding is only needed if you connect your own autonomous agent, such as an n8n workflow or a custom script.
How is the MCP server different from the testimonial API?
The API route assumes a developer writes software: create a key, send REST requests, and interpret what comes back. The MCP route assumes a conversation: the chat client authenticates for you and calls a predefined set of testimonial tools, so no code is required from you. Underneath, both approaches access the same testimonial data.
Is it safe to connect an AI agent to my testimonials?
You choose exactly what it can access at connection, from viewing walls to editing testimonials, and every tool call is bound by those permissions. Agents cannot delete a testimonial or a wall. You can revoke access from your AI client's connector settings, or revoke an autonomous agent key in the dashboard.
Which plans include AI agent access?
{{OWNER_PLAN_INCLUSION: confirm which paid plans include AI agent access, any limits, and how to upgrade}}
How is this different from the AI Testimonial Assistant?
The AI Testimonial Assistant helps customers write a better testimonial while they complete your collection form. The MCP server is for you: it lets your own AI agent read and manage testimonials you have already collected.