You can now build @activepieces flows with Claude or other MCP clients ๐ฅ
No need to think through the full workflow logic step by step.
Just connect to the Activepieces MCP server, describe the flow you want, and your AI agent can build and run it for you inside Activepieces.
You can find the MCP server in your project settings.
Check it out now and let us know what you think.
Building pieces in Activepieces just got easier ๐
We shipped 2 big updates for developers:
โ ๐ช๐ฒ ๐ฟ๐ฒ๐ฝ๏ฟฝ๏ฟฝ๏ฟฝ๐ฎ๐ฐ๐ฒ๐ฑ ๐ก๐ ๐๐ถ๐๐ต ๐ง๐๐ฟ๐ฏ๐ผ
โณ Faster builds, less setup, and easier debugging
Learn more: https://t.co/Y9W7HdL10l
โ ๐ช๐ฒ ๐ฎ๐ฑ๐ฑ๐ฒ๐ฑ ๐ฎ ๐ฃ๐ถ๐ฒ๐ฐ๐ฒ ๐๐๐ถ๐น๐ฑ๐ฒ๐ฟ ๐ฆ๐ธ๐ถ๐น๐น ๐ณ๐ผ๐ฟ ๐๐ ๐ฐ๐ผ๐ฑ๐ถ๐ป๐ด ๐ฎ๐ด๐ฒ๐ป๐๐
โณ This teaches tools like Cursor or Claude Code how Activepieces pieces work, so they can build better pieces with less guesswork
Check out the knowledge base: https://t.co/Twzoe9Cq2s
This is part of our ongoing work to make building on Activepieces faster, simpler, and more AI-friendly.
Let us know what you think.
If your AI โpilotโ has been running for monthsโฆ
Itโs not a pilot.
Itโs a bottleneck.
AI in production requires structure to scale fast.
Activepieces gives you the system to ship it.
๐ Start building:
https://t.co/f6FUO8ZpJt
An effective AI workflow isnโt magic.
Itโs structure.
โ A trigger
โ An AI step
โ Human approval
โ Logged actions
Humans decide.
AI executes.
Everything is tracked.
Thatโs how it should work.
Start here: https://t.co/4CNMTeTTsk
AI inside a workflow doesnโt just give you output.
It gives you accountability.
Because AI isnโt powerful in isolation.
Itโs powerful when it runs inside a system you can trust.
This is how companies use Activepieces
https://t.co/4CNMTeTTsk
AI without security is risk.
AI with security is control.
Activepieces connects AI agents + your work softwares
into a governed workflow.
https://t.co/f6FUO8ZpJt
A flashy AI demo gets attention.
A governed AI workflow gets adoption.
Activepieces enables you to give AI to every team with
IT control and enterprise security
https://t.co/4CNMTeTTsk
A flashy AI demo gets attention.
A governed AI workflow gets adoption.
Activepieces enables you to give AI to every team with
IT control and enterprise security
https://t.co/4CNMTeTlCM
AI adoption works when:
ย ย โข It has an owner.
ย ย โข It has approvals.
ย ย โข It has visibility.
Otherwise, itโs just another demo.
Activepieces turns AI into operational workflows.
https://t.co/4CNMTeTlCM
This flow turns every incoming support email into a proper Linear issue:
ย ย โข ย The problem is rewritten clearly
ย ย โข The description is cleaned up
ย ย โข The priority is set automatically
Template:
https://t.co/0RijmYZ3wF
New tickets shouldnโt be stuck in Zendesk until someone remembers.
This flow sends a Slack message every time a new Zendesk ticket is created, including:
ย ย โข ย Ticket subject and requester
ย ย โข Key details for quick triage
ย ย โข A direct link back to the ticket
Template: https://t.co/uWA2LKBiBe
Shipping features is fun.
Writing release notes? Not so much.
This automation:
ย ย โข Watches GitHub for new releases
ย ย โข Turns them into clean, structured release notes
ย ย โข Shares updates in Slack
ย ย โข Logs full details in Notion
Template: https://t.co/Su7Ef3usai
Most AI ROI slides are fiction because they measure intent.
Real ROI shows up as:
ย ย โขClear time and money saved
ย ย โข fewer handoffs
ย ย โข faster cycle time
ย ย โข fewer escalations
ย ย โข fewer errors
Most security teams are fine with AI.
Theyโre just not fine with unknown outcomes
The following needs to be clear for successful AI implementation:
ย ย โข What the AI can access
ย ย โข What itโs allowed to do
ย ย โข How youโll log and audit every action
Governance is what makes AI safe enough to deploy.
The fastest way to kill an AI pilot:
Tell everyone to โtry AIโ without agreeing on what success looks like.
A real pilot should be able to answer:
โข What specific task is being improved?
โข How will success be measured?
โข Who owns the outcome?
โข What safeguards are in place if something goes wrong?
If you canโt answer those, itโs not a pilot.
Itโs just experimentation with no path to results.
If your AI strategy is โgive everyone ChatGPTโโฆ you donโt have a strategy.
You have a new writing app.
Real adoption starts when AI can:
ย ย โข Pull data from systems
ย ย โข Act inside workflows
Most enterprise AI rollouts fail for a simple reason:
Using it means leaving your tools, opening a new chat window, and copy/pasting
This makes people hesitate to adopt it.
Effective AI adoption uses AI to remove extra steps to getting work done
If you can explain a task, you can automate it.
Automation isnโt only for โseriousโ workflows
Even small, random things can be automated.
Here are 6 interestingly specific things you can build with Activepieces.
https://t.co/Jc52UGuBo1
Manually planning events from Discord:
Scroll channels โ screenshot messages โ write event copy โ post.
This automation turns that into:
โNew questions in, new event drafts out.โ
CM just approves and hits go.
A few minutes of setup = hours saved every month.
Template: https://t.co/1vc05TlmEC