Pipedream

About Pipedream

Pipedream is a developer-first integration and workflow-automation platform founded in 2019 by Tod Sacerdoti and headquartered in San Francisco. It targets the gap between no-code tools that cannot express arbitrary logic and bespoke integration code that nobody wants to host. A Pipedream workflow is a sequence of steps triggered by an event source, where each step is either a pre-built action or a block of code the developer writes inline.

The defining feature is first-class code. Inside any workflow step a developer can write Node.js, Python, Go, or Bash, with npm and PyPI packages available by import and no separate build or deploy stage. This makes Pipedream closer to a hosted serverless runtime with an integration layer on top than to a visual builder with an "escape hatch" code block. The platform connects to more than 2,700 applications and maintains managed authentication for them, so a workflow can call an authenticated API without the developer handling OAuth token storage or refresh.

Pipedream pricing as of May 2026 is metered in credits, where one credit equals 30 seconds of compute at 256 MB of memory and credit cost scales with memory (512 MB doubles it, 1 GB quadruples it). The free plan includes 100 credits per day with hard caps that stop a workflow mid-run when exceeded. The Basic plan is $29 per month for 2,000 credits per day, and the Advanced plan is $79 per month for 10,000 credits per day, longer event-history retention, and custom domains for HTTP endpoints. The Business plan is custom-priced and adds unlimited credits, single sign-on, audit logging, and service-level agreements.

flowchart LR
  A[Event source: webhook, schedule, app trigger] --> B[Pipedream Workflow]
  B --> C[Pre-built action step]
  B --> D[Code step: Node.js / Python / Go / Bash]
  C --> E[Managed auth + connected app APIs]
  D --> E
  D --> F[npm / PyPI packages]
  E --> G[Result / next step / HTTP response]

Beyond standalone workflows, Pipedream has expanded into embedded integrations. Pipedream Connect lets a SaaS company offer its own users a library of third-party integrations without building each one, using Pipedream's connector catalogue and managed authentication behind the customer's own interface. The company has also added Model Context Protocol support, exposing its connector catalogue to AI agents in the same way it exposes them to workflows.

Pipedream competes with Zapier, Make, n8n, and Workato, and with serverless platforms such as AWS Lambda for the code-heavy end of the spectrum. Its niche is well defined: teams that have developers, want managed authentication and hosting, and refuse to be limited by a visual builder's logic constraints. It is a weaker fit for non-technical operators, who are better served by Zapier, and for teams that need to keep all execution inside their own infrastructure, who are better served by self-hosted n8n or Activepieces.

Editor's Note: Across 7 ShadowGen Pipedream builds in 2024-26, the credit model is honest but easy to misjudge. The mistake we see most: a workflow that calls a slow upstream API sits idle waiting on the response and still burns compute credits the whole time. In one client integration a single 9-second third-party API call, run a few thousand times a day, quietly consumed most of a Basic plan's daily credits. The fix that works: move slow or unreliable upstream calls behind a queue or a separate async workflow rather than blocking inside the main run. For fast API work, the free and Basic tiers stretch a long way. — Rafal Fila, ShadowGen