Pipefork is a browser-based API builder. Combine upstream APIs, shape the response, and publish a secure endpoint. Start with AI Builder or manage your APIs through MCP.
Free plan for development and evaluation. No credit card required. Access is currently by private beta invitation.
Customer overview
GET /customer-overview/{customerId}
Request
GET /customer-overview/{id}
HTTP step
GET crm.example.com/customers/{id}
HTTP step
GET billing.example.com/invoices
Transform
Compose clean JSON response
AI Builder
✓Endpoint draft created
✓Upstream calls configured
✓Response mapping proposed
✓Ready to review and test
Build a useful endpoint from the services you already use, then manage it from draft to release.
Call multiple HTTP APIs and shape their results into one JSON response.
Describe an endpoint and review the draft AI Builder creates.
Let authorized AI clients help manage projects, drafts and deployments.
Protect endpoints with API keys and promote versions through Dev, Test and Prod.
Start in the browser, with AI Builder, or through an authorized MCP client. Each path creates a Pipefork API draft you can review and deploy.
Start manually or tell AI Builder what the endpoint should do.
Connect the HTTP APIs you need and pass data between steps.
Return the fields your clients need while keeping upstream credentials private.
Test the draft, deploy to Dev, and promote a version through Test and Prod.
Model Context Protocol (MCP) connects AI tools to Pipefork. Use Management MCP to manage your workspace, or a published API's MCP endpoint to call its deployed tools.
Connect an authorized MCP client to create and update projects, prepare drafts, and manage deployments.
Expose selected deployed endpoints as tools an AI client can call.
First, copy your Management MCP server URL from Pipefork. Add it to your AI tool, then sign in to approve access to your workspace.
Work with your Pipefork workspace from ChatGPT.
Availability depends on your ChatGPT plan and workspace settings.
Official setup guideGive Cursor Agent access to your Pipefork tools.
Cursor may need Pipefork’s optional OAuth dynamic registration enabled by the deployment operator.
Official setup guideConnect Claude to Pipefork through a custom connector.
On Team and Enterprise plans, an owner adds the connector first.
Official setup guideUse any compatible remote MCP client.
Your client must support remote MCP, OAuth, and a registration method enabled by Pipefork.
Official setup guideThis connection manages Pipefork. To let an AI client call a published API, use that API's own MCP URL from its published documentation.
For developers and small teams building frontend backends, partner integrations, internal tools, and AI agent workflows.
Explore Pipefork on Free. Paid plans add production environments and higher limits.
Free
For evaluation and development only. No production commitment.
Free
No credit card required
Starter
For real projects running in production with Dev, Test, and Prod.
TBA
Coming at launch
Pro
Team-scale plan for larger API portfolios and higher limits.
TBA
Coming at launch
Sign in
Sign in with your email address