I’m using Diaflow for end-to-end content automation to publish posts on a Facebook Page, and it’s honestly one of the most enjoyable workflow/automation tools I’ve worked with.
My real-world use case

My workflow is close to 30 nodes, but most of those are non-AI automation steps and are effectively “free” in terms of credits. The real credit usage comes from AI (I use Anthropic, so token cost is the main driver). My pipeline looks like this:
Pull content signals from email
Download an attached PDF
Send it through AI to extract/transform and write the post
Send the drafted post to Telegram (quick review/visibility)
Store it in a database
Publish to Facebook Page
Log the results to Google Sheets
On average, each post costs about 40–50 credits, and in my case that’s mostly due to AI/token usage — not the basic workflow nodes.

What feels truly worth it
Outstanding support: responsive, thoughtful, and genuinely helpful with real production workflows.
Fast product iteration: the web experience has improved dramatically compared to the early days. You can feel the team is shipping continuously.
Polished UX: building and maintaining larger flows is clear and manageable.

Constructive feedback (still 5 stars)
If Diaflow wants to be even more power-user friendly, I’d love to see:
More clarity and optimization around credits for basic actions (especially for “obvious utilities” like requests/logging/updates), while keeping AI cost transparent.
A clearer cost estimate before execution (expected credits/tokens), so users can predict spend with confidence.

Bottom line
If you’re building content automation — especially pipelines that include AI processing + publishing + logging — Diaflow is absolutely worth it. This isn’t a “cheapest possible hobby tool”; it’s a serious workflow platform aiming for reliability, UX, and strong support. For me, it’s an easy 5 stars.

On pricing/credits: I understand why the Founder/team structured it this way. Workflows that involve AI, parsing PDFs, API calls, retries, deliverability constraints, and reliability guarantees all come with real and sometimes unpredictable costs. If the product is built for business users who care about stability and compliance, a credit model that protects quality and prevents abuse can be perfectly reasonable.
That said, I do hope the team continues improving transparency around “what actually costs money” and further optimizes credits for basic non-AI actions so power users don’t feel nickel-and-dimed. And for users expecting near-zero pricing while running complex workflows, it’s worth being fair: this is a serious automation platform, not an unlimited free utility. Feedback grounded in real usage and operational realities helps the product improve far more than purely emotional complaints.