Pipedream has launched mcp.pipedream.com with dedicated MCP (Model Context Protocol) servers for 2,500+ integrated apps.
These servers enable AI assistants like Claude and Cursor to securely access and interact with 1000s of APIs through a standard protocol, or deploy Pipedream MCP servers to your agent and add 2,500+ APIs with 10,000+ tools in minutes.
Pipedream handles the auth and tooling so you can focus on shipping.
Not because it isn’t useful (to be honest it looks decidedly useful as an API aggregator service) but the fact that one of its main selling points is for “AI”.
Yet, if we had AI, it would be able to build these integrations itself
AI building the integrations is the approach we are taking at the moment, to the extent that you can store credentials like API keys and Oauth and then make the AI aware of these via a variable like ${googleAPI.SECRET}.
These credentials can then be use in the API node (where the AI can write a custom curl) or the run code component (where the AI can write custom python or js code to make an HTTP request).
Can it automatically update themselves based on changes to the source API? Can it detect ahead of time if APIs have changed? How about understanding the API requirements?
I encourage anyone who is curious to try to build an AI to get an approved client ID / secret from Google to access end user's gmail accounts. Most tier-one apps, such as Google, Zoom, Shopify, etc require a formal application process, third-party security reviews, and much more.
I think my point went over the head here, so to speak.
The fact is “AI” isn’t building these integrations because they aren’t capable. They’re capable of generating some sandbox code, if anything, but full scale integration, that updates itself as APIs change etc?
It’s simply not happening, because they don’t actually reason, because they lack actual intelligence
Honestly I thought it was a joke or /s website at first. I looked at the site/docs and still don’t understand what your service is exactly selling/doing?
Pipedream co-founder here. We excited to introduce Pipedream Connect, which is the easiest way for your users to connect to over 2,400+ APIs, right in app or AI agent.
Connect provides managed authentication, approved client IDs, durable components, and infrastructure for running serverless functions. Build in-app messaging, CRM syncs, AI agents, and more in minutes. Key features:
- Handle authorization or accept API keys on behalf of your users, for any of Pipedream's 2,400+ APIs. Use the Client SDK or Connect Link to accept auth in minutes.
- Securely retrieve OAuth access tokens, API keys, and other credentials for your end users with Pipedream's REST API
- Run workflows for your end users with Pipedream's workflow builder, serverless runtime, and thousands of no-code triggers and actions. Build complex integrations in minutes, writing code when you need it and using no-code components when you don't.
- Run any Pipedream action or deploy any Pipedream trigger on behalf of your users, directly from within your application.
Today, managed authentication is publicly available.
Pneumatic looks cool but you are clearly missing the necessary app integrations to be competitive. If you want to integrate 2200+ applications in your product (with 10k+ triggers and actions) instead of sending all your customers to competitive products like Zapier and Make.com, you should explore Pipedream Connect - https://pipedream.com/connect.
Pipedream Connect is the easiest way for your users to connect to any API,
right in your product.
Connect provides managed OAuth tokens, a library of API helpers, and a platform for running serverless functions for any of your users. Build features that connect to 2,000+ APIs in minutes.
Pipedream Connect is currently in preview, and we're looking for feedback
During the Preview phase, Connect is free to use for any workspace. Pricing will come later and will be clearly communicated in advance.
