Quick answer: the best Pipedream alternative for each need
Want open source you can self-host? Look at Activepieces or Windmill. For background tasks and durable functions written as code, look at Trigger.dev, Inngest or Temporal Cloud. Replacing Pipedream Connect for agent tool calling and managed user auth points at Composio or Nango. For a low-code builder with code steps, look at Latenode, Make or Zapier.
Every price and limit below was read from the vendor's own pricing page in September 2026. Where a tier is custom-priced, it is described by its billing model rather than by a number someone guessed. No reseller or aggregator figures were used. The ten tools are grouped by the job they are built for, not ranked against each other. A durable execution engine and a visual builder are not competing for the same task. The sections run in the same order as the comparison table.
Why developers look for Pipedream alternatives in 2026
Pipedream's Free plan costs $0 a month with no credit card required. It includes up to 100 credits a month, 1M AI tokens, 3 active workflows and 3 connected accounts. Basic costs $29 a month, billed annually as $348, with 2,000 credits a month, 20M AI tokens, 10 active workflows and 5 connected accounts. Advanced costs $49 a month, billed annually as $588, and adds 50M AI tokens, control flow operators, unlimited workflows, unlimited connected accounts, premium apps and GitHub Sync. Connect costs $99 a month, billed annually as $1,188, with 10,000 credits a month and auth for 100 external users. Business is custom-priced with volume pricing, support for HIPAA workloads and uptime SLAs. All prices are listed in US dollars.
Why teams shop around is visible in that list. The meter is credits, so forecasting spend means forecasting credit consumption rather than run time or seats. Limits on the cheaper plans sit on objects as well as on volume: 3 active workflows and 3 connected accounts on Free, 10 and 5 on Basic. That is easy to outgrow when every client integration needs its own account. Control flow operators arrive only at Advanced. And Connect is priced per external user, so the bill tracks your customer count rather than the work actually done.
None of that makes it a weak product. Pipedream integrates 3,000+ apps and lets workflow steps run Node.js, Python, Go and Bash, a wider language set than the visual builders on this list offer. That code-first model is why people picked it. The alternatives below compete on a different axis: self-hosting, durability guarantees, a cheaper meter at volume, or auth for your users' accounts.
Pipedream alternatives at a glance
Ten tools here. The job each one is built for, what its meter counts, and what you get before paying anything. Plan names, rates and caps behind each row sit in the sections that follow.
| Tool | Best for | Pricing model | Free tier |
|---|---|---|---|
| Latenode | Code steps by runtime | CPU seconds | 10,000 CPU seconds |
| Activepieces | Open-source low-code | Credits | 100 credits a day |
| Windmill | Self-hosted scripts | Seats at Enterprise | Community Edition |
| Trigger.dev | Background tasks | Compute plus runs | $5 credits a month |
| Inngest | Durable functions | Executions | 50k executions |
| Temporal Cloud | Critical orchestration | Actions | $1,000 free credits |
| Composio | Agent tools and auth | Tool calls | 100,000 tool calls |
| Nango | In-product API auth | Connections | 10 connections |
| Make | Visual scenarios | Credits | 1,000 credits |
| Zapier | Widest app catalog | Tasks | 100 tasks a month |
1. Latenode: best for code steps with runtime billing
Latenode is a hosted low-code builder. You assemble nodes on a canvas and drop into custom JavaScript when a visual step falls short. Every workflow gets an AI Copilot, and the platform adds an AI agent builder, a headless browser for sites with no API and 335+ LLM models.
Billing is the real difference. Latenode charges for workflow runtime in CPU seconds, not per operation and not per step, so extra steps do not multiply the bill. The first 10,000 CPU seconds are free every month on both plans. Free is $0 a month with no credit card, 5 active workflows, one execution worker, 3 minutes per run and 3 days of history. Pay as you go adds no base fee and includes the same 10,000 seconds. From 10,001 to 100,000 it charges $0.00012 per CPU second, with lower rates above that. Caps lift to unlimited scenarios, 5 parallel workers, 10 minutes per run and 30 days of history.
The trade-off: this is a hosted platform, not an open-source engine you run yourself. A runtime meter is not comparable to tasks or credits. Estimate with average run duration.
2. Activepieces: best open-source Pipedream alternative for low-code automation
Activepieces is the closest thing here to a drop-in swap for a Pipedream builder workflow. It is open source at the core, offers 760+ integrations, and can be embedded inside another product. That matters if your own customers use the automation. Every plan runs on Activepieces' cloud or on your own servers at the same price, so self-hosting is a deployment choice, not a different SKU.
The Free plan costs $0 with 100 credits a day for one user. Plus is $16 a month billed annually with 10,000 credits a month for up to 5 users. Team is $166 a month with 50,000 credits for 25 users, and Ultimate is custom pricing. Overage on Plus and Team costs $0.007 per credit. One credit covers a whole run for ordinary steps regardless of how many there are, which is friendlier than a per-module meter for long flows.
The catch is AI. Steps that call a model cost more than one credit: 2 for a fast model, 10 for a smart model and 20 for a frontier model, per step. An AI-heavy flow spends a monthly allowance faster than the per-run number suggests. Count model steps separately when sizing a plan.
3. Windmill: best free self-hosted platform for scripts, flows and internal apps
Windmill turns scripts into flows and internal apps. That fits teams whose automation is really glue code plus a small UI for someone to click. The Community Edition is free and open source to self-host, with unlimited executions, up to 10 users with SSO and 100 free app guests per 30 days. For a small platform team that combination is unusual: no execution meter, and guests can use an app without a seat.
Paid value sits in the Enterprise edition, which was listed in September 2026 from $120 a month and priced per seat. Developer seats run $20 a month each, operators $10 a month each. It adds audit logs, SAML with group sync, uncapped SSO, a commercial license and 24/7 priority support with a 3-hour response time.
Two limitations to plan around. Enterprise is seat-priced, so cost grows with the size of your team rather than with how much you run. That inverts the logic of every usage meter in this article. And the hosted cloud plan is quoted through a calculator rather than a list price, so check the vendor's pricing page before you budget for a managed deployment.
4. Trigger.dev: best open-source background tasks with usage billing
Trigger.dev is built for background work that outlives a request: long jobs, retries, queues and scheduled tasks you write and version alongside your application. It is open source under the Apache 2.0 license and can be self-hosted. Tasks run with no timeouts, which removes the usual serverless ceiling on slow work.
The Free plan is $0 a month and includes $5 a month in free credits, 20 concurrent runs, unlimited tasks and 5 team members. It also covers dev and prod environments and 1-day log retention. Hobby is $10 a month with $10 of credits included and 50 concurrent runs. Pro is $50 a month with $50 of credits and 200 concurrent runs, with extra concurrency at $10 a month per 50 runs. Enterprise is custom. Usage is charged two ways: seconds of compute per machine size, plus an invocation cost for each run that starts on managed workers. Runs in the dev environment are not charged. Trigger.dev's own worked example puts a 10-second task running 100 times a day at about $1.09 a month on the small-1x machine.
The limitation is concurrency, not volume. 20 concurrent runs on Free is a real ceiling for spiky workloads, and extra concurrency is sold on Pro in blocks of 50 runs.
5. Inngest: best durable execution for event-driven functions
Inngest positions itself as open source durable execution. You wrap existing code in step.run(), and each step handles retries, state and failure recovery instead of you writing that machinery again. It runs locally through a dev server started with npx inngest-cli dev, so the same functions you deploy are the ones you debug on your laptop. For event-driven work with fan-out, waits and retries, this is a smaller conceptual jump than moving to a visual builder.
"An execution is a single durable function run plus each step inside it. If your function has 5 step.run() calls, one run uses 6 executions (the run itself + 5 steps)."
Source: inngest.com
Hobby costs $0 a month with no credit card and includes 50k executions a month. Pro starts at $99 a month with 1M executions included and pay-as-you-go usage up to 20M. Enterprise is custom.
Read the meter carefully before you compare that with anything else. An execution is a single durable function run plus each step inside it, so a function with 5 step.run() calls uses 6 executions per run. A step-heavy function is exactly the kind you want durability for. It is also the kind that burns the Hobby allowance faster than a headline of 50k suggests. Count the steps in your busiest function, multiply, then look at the number again.
6. Temporal Cloud: best for mission-critical workflow orchestration
Temporal Cloud is the option you pick when a failed workflow is a business incident rather than a missed sync. Workflows are written as code with open-source SDKs. The platform bills Actions, which represent billable operations between an application and Temporal Cloud, with additional actions priced per million.
The tiers listed in September 2026 start with Essentials from $100 a month, including 1M actions and 1 GB of active storage. Business starts at $500 a month with 2.5M actions and 2.5 GB. Enterprise is priced by sales with 10M actions and 10 GB. New accounts get $1,000 in free credits. The startup program offers $6,000 in free credits to startups with less than $30M in funding, enough runway to model real traffic before committing.
The limitation is shape, not price. The pricing page lists actions, storage and support tiers, not an app connector catalog, so every integration is code you write and maintain. There is also no permanently free plan in those tiers: the credits are a starting balance, and the entry commitment begins at $100 a month. Teams that want connectors out of the box should look elsewhere.
7. Composio: best for agent tool calling and managed auth
If you used Pipedream Connect for agent tools and users' OAuth tokens, Composio is the closest replacement. Every plan includes 1500+ toolkits, OAuth management, triggers, sessions, custom tools and MCP. The auth handshake and tool schemas come from the platform, while agent logic stays in your code. Each tool run counts as one tool call, and meta tools like tool search are free.
"Free is hard-capped, needs no credit card, and never produces a surprise bill. When you reach a cap, usage pauses until the next month or until you upgrade."
Source: composio.dev
Free costs $0 with 100,000 tool calls and 50,000 trigger events a month, unlimited connected accounts on your own app and 3 team members, no credit card required. It is hard-capped: usage pauses until next month or an upgrade. Pro costs $29 a month with a $29 monthly usage credit that expires at the end of each billing month, plus unlimited team members and spend controls. Usage continues at overage rates of $0.0003 per tool call and $0.003 per trigger event. Enterprise is custom, adding volume discounts, KMS, SSO, SCIM, MSA, DPA and SLA.
Watch the shared-app limits listed in September 2026. On Composio-managed apps, up to 20K of the 100K free tool calls and 1K connections are free, then $0.10 per connection on paid plans. Log retention is 7 days on Free, 30 on Pro.
8. Nango: best for API auth and syncs inside your own product
Nango solves the other half of the Connect problem: authenticating your users against third-party APIs and keeping data in sync, inside a product you ship. It provides authentication with 900+ APIs and MCPs plus 6k pre-built tools, triggers and syncs. A new integration is configuration rather than another OAuth implementation.
The meter is the connection, and it maps to reality cleanly. One connection equals one API key or access token of an external API, in other words one authorized user account, billed per connection per month, prorated on paid plans. The Free plan costs $0 with hard-capped limits that reset monthly: 10 connections, 10 hours of compute time a month and 10 GB of data transfer a month. Pay-as-you-go costs $50 a month and includes $50 in credits each month, drawn down at $0.29 per connection, $0.72 per compute hour and $0.5 per GB. Enterprise is custom and adds self-hosting, an MSA and an SLA.
The limitation is inherent to that model. Ten connections is a prototype allowance. Past it your bill grows with your customer count, the same dynamic that drives the Connect bill. The difference is the unit price, and that syncs and compute are metered separately.
9. Make: best visual builder with code on paid plans
Make is the visual option for people who think in diagrams. Scenarios are drawn as connected modules, and every plan includes multi-step workflows and instant webhook execution. On paid plans the Make Code app supports custom Python and JavaScript, and Enterprise adds Custom Functions applied inline. Make MCP Server turns scenarios into callable tools for ChatGPT, Claude, Cursor and other MCP-compatible agents, covering 3,000+ apps and 30,000+ actions. It is a straightforward way to expose existing automations.
The Free plan costs $0 a month with up to 1,000 credits per month, and scheduled scenarios run no more often than every 15 minutes. Core adds scheduling down to the minute. Billed annually, Core is $9 a month, Pro $16 and Teams $29, each with 10,000 credits a month.
Model the meter first. A credit is consumed by each module action in a scenario, such as adding a Google Sheet row or fetching Gmail data. Each module that processes data typically consumes at least one credit, and AI or multi-bundle features consume more. Code execution is billed at 2 credits per second. A long scenario therefore costs more per run than a short one doing the same job, the opposite of a per-run meter.
10. Zapier: best app catalog with Code steps
Zapier remains the widest catalog: more than 9,000 app integrations by its own count. API updates and connector maintenance are handled in-house, the practical reason non-engineers can keep a Zap running. Code steps let you add a Python or JavaScript step inside a Zap when a connector stops short.
The Free plan costs $0 a month and includes 100 tasks per month with Zap workflows, Tables and Forms. Professional costs $19.99 per month billed annually and includes 750 tasks a month, or $29.99 per month when billed monthly. The advertised discount for paying yearly instead of monthly is 33%. Enterprise pricing is not published and requires contacting sales.
A task is anything the platform successfully completes on your behalf, such as sending an email or creating a spreadsheet row. Triggers, filters and failed steps are free, so polling that finds nothing costs nothing and only completed work counts. Task usage is pooled across the entire account, so Zap workflows, AI steps, code, MCP and SDK all draw from the same monthly allocation. One Zapier MCP tool call consumes two tasks. Pooling is convenient until volume arrives. At scale the per-task price, not the app count, decides whether you stay.
How to choose a Pipedream alternative
Work through these in order. The first two eliminate half the list.
- Where the code runs. Activepieces, Windmill, Trigger.dev and Inngest run on your own servers; Latenode, Make, Zapier, Composio and Temporal Cloud are hosted; Nango self-hosts on Enterprise.
- What the meter counts. Credits on Activepieces and Make, executions on Inngest, tool calls on Composio, connections on Nango, actions on Temporal Cloud, CPU seconds on Latenode, tasks on Zapier, compute plus invocations on Trigger.dev. Map your busiest workflow onto each unit.
- Free-tier caps against what you actually run. Count your live workflows and connected accounts, then check the numbers above: object caps bite before volume caps.
- Whether you need auth for your users' accounts. That is the Connect use case, and it points at Composio or Nango, not at a builder.
- Durability. If a half-finished run is unacceptable, you want retries, state and failure recovery as platform features: Inngest, Temporal Cloud and Trigger.dev are built that way.
- Log retention. One day on Trigger.dev Free, 3 days of run history on Latenode Free, 7 days on Composio Free and 30 on Pro. Short retention turns a Friday incident into guesswork.
- Team seats. Activepieces allows one user on Free, 5 on Plus and 25 on Team; Windmill's Community Edition covers up to 10 users; Trigger.dev Free includes 5 team members, Composio Free 3.
If you arrived from a search for free or open source, separate the two. Windmill's Community Edition, Activepieces, Trigger.dev under Apache 2.0 and Inngest run on infrastructure you control: free of license cost, not free of operations work. Latenode, Make, Zapier, Composio and Nango offer free tiers on a hosted service: nothing to run, but a cap you will meet. Temporal Cloud offers credits rather than a free tier. Pick the free that matches who maintains it at 2am.
