AI & Machine Learning · PromptHub + Claude
PromptHub MCP with Claude: Versioned Prompts Your Team Can Trust
The short answer
Claude works with PromptHub through the Model Context Protocol (MCP), an open standard that acts as a secure adapter between the two. Once connected with an API key, Claude can retrieve stored prompts, run specific versions, compare revisions and update prompt metadata, all through plain-language requests. Your team sets the permissions.
What is PromptHub?
PromptHub is a prompt management platform. It stores the instructions, or prompts, that power AI features, tracks every version, and lets software call those prompts through an API. Engineering, product and operations teams use it to ship AI features reliably, keep prompt changes traceable, and reuse proven prompt assets across projects instead of rewriting them.
How Claude connects to PromptHub through MCP
The Model Context Protocol is an open standard that lets Claude securely use other software, reading information and taking actions, with permissions the organisation controls. Think of it as a universal adapter between Claude and the tool. Claude does not need a custom integration for each product.
Once PromptHub is connected using an API key, Claude can typically list the prompts your team has stored, retrieve a chosen version, run it with supplied inputs, compare two revisions side by side and read or update prompt metadata such as names, tags and descriptions. You ask in plain English, and Claude carries out the steps.
Permissions and approvals keep people in charge. The API key is scoped to the access your team decides on, so Claude only sees what it needs. Consequential actions, such as publishing a new prompt version, can be set to require a person's approval before they take effect, and activity is logged for review.
What you can automate
Finding the right prompt quickly
A product manager asks Claude to find every prompt tagged for customer support summaries and show the latest version of each. Claude retrieves the matches and lists them with their descriptions, saving the search through folders and old documents. The team spends less time hunting and more time building.
Comparing prompt revisions before release
An engineering lead asks Claude to compare version 14 and version 15 of a classification prompt. Claude highlights the wording changes and summarises what each edit is likely to do. Reviewers can then decide which revision to ship with a clear, documented reason.
Testing a prompt against sample inputs
A quality analyst asks Claude to run a chosen prompt version against a set of twenty sample customer messages and report the outputs. Claude returns the results in a table for review. This gives the team an early view of how a change behaves before it reaches real users.
Keeping prompt documentation current
A knowledge manager asks Claude to review prompt metadata and flag any prompts missing an owner, a description or a tag. Claude produces a list, and the manager approves the updates. Prompt libraries stay organised as they grow.
Reusing proven prompts across teams
A department head in operations asks Claude to find prompts already written for invoice data extraction, so a new team does not start from scratch. Claude shows the options and their version history. Proven work is reused rather than duplicated.
An end-to-end workflow: Governed prompt release for a customer support assistant
TriggerA support lead asks Claude to prepare a new version of the reply-drafting prompt after reviewing recent customer complaints.
Claude retrieves the current production version of the reply-drafting prompt from PromptHub.
Claude drafts a revised version based on the complaint themes and the lead's instructions.
Claude runs both versions against a set of past support tickets stored in the helpdesk system and compares the outputs.
Claude sends a summary of the differences and quality observations to the support lead in Microsoft 365 for review.
After the lead approves, the revised version is saved in PromptHub with a change note, and the audit log records who approved it.
OutcomePrompt changes reach customers through a reviewed, documented process, with every version traceable. Typically this shortens the review cycle and reduces untracked edits to live AI features.
How we set it up
Book a free automation audit
We review how your teams store, test and release prompts today and identify the workflows worth connecting to Claude. There is no cost and no obligation.
Set up secure access
We create a PromptHub API key scoped to the least access the workflow needs, and store it in your secure secrets manager. Claude connects through MCP using that key.
Design the end-to-end workflow
We map how PromptHub links with your other systems, such as Microsoft 365, your helpdesk or CRM, and decide which actions need a person's approval.
Test and launch with monitoring
We test the workflow against real past cases, then launch with activity logs and monitoring so you can see every action Claude takes.
Train your team
We train your staff on how to request prompt tasks in plain language, how to review Claude's work and how to escalate issues, so the team runs the workflow confidently.
Who benefits
- Engineering leads
- Ship prompt changes with a traceable history and faster, evidence-based comparisons between revisions.
- Product managers
- Find, test and reuse prompts without waiting for a developer to search for them.
- Quality and compliance teams
- Review prompt changes against a clear audit trail before they reach customers.
- Operations and department heads
- Reuse proven prompt assets across teams and reduce duplicated effort.
Security and control
- Scope the API key to only the projects and actions the workflow needs, and rotate it on a regular schedule.
- Require human approval before any new prompt version is published to live systems.
- Keep activity logs of every retrieval, run and metadata change for audit review.
- Test workflows against real past cases before launch, and monitor them afterwards.
- Store credentials in a managed secrets vault, never in shared documents or chat messages.
For teams in the US, UK, India and the Middle East
Organisations in the United States and the United Kingdom can host the Claude workflow in their chosen region, supporting their data protection obligations. In India, the Digital Personal Data Protection Act, 2023 shapes how personal data in prompts is handled, and we design around it. For organisations in the Middle East, including the UAE and Saudi Arabia, we follow the relevant national data protection laws and select a hosting region that meets those requirements.
Learn more about Claude and MCP
Anthropic's AI model and assistant.
Secure links between Claude and business tools.
How engineers build and change software with Claude.
Hosted agents that run multi-step work.
The open standard Claude uses to connect to tools.
Frequently asked questions
What does the PromptHub MCP server do?
The PromptHub MCP server acts as a secure bridge between Claude and your PromptHub account. It lets Claude retrieve stored prompts, run versions, compare revisions and manage prompt metadata when you ask it to. The server only does what its permissions allow, so your team controls exactly which prompts and actions are available.
Is connecting PromptHub to Claude secure?
Security depends on how the connection is set up. We use an API key with the narrowest access the workflow needs, store it in a managed secrets vault, and log every action. Consequential changes can require human approval. We also test the setup before launch so you can see exactly what Claude can access.
What can Claude do with PromptHub once connected?
Typically Claude can find prompts by name or tag, retrieve a specific version, run it against sample inputs, compare two revisions and read or update descriptions and other metadata. You give these instructions in everyday language, and Claude carries out the steps within the permissions you have set.
How long does setup take?
Timing depends on how many systems are involved and how complex the workflow is. A focused connection with one clear workflow often takes a few weeks from the free audit to launch and training. We agree a timeline during the audit so you know what to expect before committing.
Which PromptHub plan or permissions do I need?
You need a PromptHub account that allows API access, and an API key with permissions matching the workflow you want. The exact plan requirements depend on PromptHub's current terms, so we confirm what your account supports during the free automation audit and recommend the least-privilege setup.