Backendless

Your automation works. But the data coming through looks terrible. Raw database IDs showing up in user notifications. ISO timestamps when you need “2 days ago.” Arrays that need filtering before you can use them. Every flow ends up with extra data-munging steps just to make things presentable. The Transform Data block handles this in one place: - Map and restructure objects - Extract substrings or check for keywords - Flatten or filter arrays - Reformat dates or calculate time offsets - Run inline logic like conditionals or switches It uses the Expression Editor, so you can combine data from previous blocks, constants, or computed values on the fly. Instead of passing messy data downstream and dealing with it in every subsequent block, you transform it once. Clean input means simpler logic everywhere else. No more extra flows just to parse basic data. No more raw IDs leaking into customer-facing messages. No more timestamp mismatches breaking your integrations. Transform once, use everywhere. Learn more: https://lnkd.in/gNyn9B95 Try it: https://www.flowrunner.ai #aiagents #aiautomation #automation #workflowautomation #flowrunner

Need to respond to webhooks or expose flows as API endpoints? That’s what Return Result is for. When external systems call your flow - like Twilio webhooks, Stripe payment notifications, or custom API integrations - they expect a structured response back. Without Return Result, you can’t control what gets returned. The Return Result block lets you: - Define the exact response structure for webhook calls - Return properly formatted data to external services - Meet API contract requirements (like TwiML for Twilio) Real example: Twilio sends an SMS webhook to your flow. You process the message, look up the customer, log it to your database. Twilio expects specific XML or JSON back to confirm receipt or send a response. Return Result lets you construct exactly that response. Use it when: - Handling webhooks from external services - Exposing flows as callable API endpoints - Responding to external system requests with structured data Return Result is a terminal block - it ends execution and sends your defined response back to whatever called the flow. It’s the difference between flows that just process data internally versus flows that integrate with external systems as proper API services. 👉 Full article here: https://lnkd.in/gfyW9H-K 🤖 Try it for yourself: www.flowrunner.ai/ #automation #flowrunner #workflowdesign #cleancode #cleanoutputs #aiagent

Parallel execution is powerful. Until downstream blocks run multiple times because they’re receiving input from multiple branches. Flowrunner automatically creates parallel execution paths when a block has multiple successors. That’s useful for concurrent processing, but it means any block downstream of the split will execute once for each incoming branch. 🔁 The Synchronize block merges parallel branches back into a single execution path. When you need multiple parallel operations to complete before continuing - API calls, validations, data processing - Synchronize waits for all incoming branches to arrive, then continues with a single execution. Configure it with: - Max waiting time (static or dynamic via Expression Editor) - Timeout behavior for branches that don’t complete in time Use it when: - Running parallel API calls that must all complete before proceeding - Coordinating multiple validation checks - Processing data batches in parallel before consolidation - Implementing approval workflows where all approvers must respond Without Synchronize, your downstream blocks execute multiple times. With it, you control exactly when and how parallel branches merge back together. Alternative: For simpler cases, Action Group with “On Completion” mode provides similar coordination without explicit synchronization blocks. Learn more: https://lnkd.in/diKNZJi5 Try it: www.flowrunner.ai/ #aiagent #aiautomation #workflow #automation

A lot of automation comes down to making a choice. Do we approve or reject? Route it to Team A or Team B? Run the fast AI model or the accurate one? You could build a long chain of if-else conditions. Or, you could just use an AI Router block. You give it a prompt (what it needs to decide), some context (like data from earlier steps), and the set of valid outcomes. It reads the input, picks one of your options, and routes the flow in that direction. That’s it. It works great for stuff like: - Classifying leads by score - Sorting support tickets by type - Deciding which model to use based on input - Routing content to “Publish,” “Flag,” or “Reject” And it always has a fallback path, so your flow doesn’t break if nothing matches. One block. Clean logic. No condition sprawl. If your flow needs to make decisions, give it the ability to choose — not just follow. 👉 Full breakdown here: https://lnkd.in/gUDPAWQF #automation #flowrunner #aiagent #workflowdesign #llmops

Sometimes, text isn’t enough. You need your automation to speak. With Flowrunner’s Text-to-Speech block, you can turn any text into clean, natural audio. Just drop it into your flow, pass it a string, and it talks. Pick the voice. Set the speed. Choose the format (MP3, WAV, FLAC; whatever you need). That’s it:) It’s perfect for: - Chatbots or phone systems that need to talk back - Smart devices that need to say things out loud - Creating audio versions of blog posts or reports - Making your flows more accessible, or more hands-free No new tools. No extra setup. Just one block — and suddenly your automation has a voice. 👉 Full breakdown here: https://lnkd.in/gXTxNkJ4 #flowrunner #automation #texttospeech #voicefirst #accessibility #builders

🔈 Audio is everywhere. But most workflows can’t do anything with it. Calls. Meetings. Voice notes. Tons of insight locked in a format that’s hard to search, store, or use. Flowrunner’s Speech-to-Text block changes that. Just drop in an audio file (MP3, WAV, M4A), and it returns clean, structured text powered by OpenAI’s Whisper model. ✅ Works in 90+ languages ✅ Handles background noise and accents ✅ Converts voice into something your automations can actually use What this unlocks: - Support calls that auto-log into your CRM - Meetings that become searchable, actionable notes - Podcasts or webinars with instant transcripts - Compliance logs you don’t have to transcribe by hand No more dead-end audio. Just structured data you can route, analyze, and build on. 👉 If your workflows touch audio, make sure they can listen. Full breakdown here: https://lnkd.in/gV3d554a #automation #audiototext #speechrecognition #flowrunner #AItools

As workflows grow, so does the need for structure. It’s not just about getting things to work. It’s about making them easy to maintain, scale, and reuse. That’s where Subflows in Flowrunner come in; Instead of building one long, fragile chain of steps, Subflows let you turn logic into modular, self-contained blocks: Build it once. Reuse it across any flow. Update it in one place. They help you: - Keep workflows readable and organized - Standardize logic across teams - Test and debug with less guesswork Whether it’s formatting data, sending alerts, or scoring leads, Subflows make it easier to do complex things simply. If your workflow feels like it’s getting heavy, try breaking it down. You’ll move faster with less friction. 👉 Read the full breakdown here: https://lnkd.in/gewaH4h3 #automation #nocode #workflowdesign #ops #flowrunner