Tools & Workflows · Methodology & funding

Rebuild it yourself. Know the limits.

Request → task · English edition of 25 September 2026 · German text as of 25 September 2026

A workflow download should deliver more than a plausible picture. That is why we keep validation rules, expected outputs and actual execution apart. The limit of each piece of evidence is part of the result.

What counts in the comparison

Task outcome: are the ID and content of the created task correct?
Failure behaviour: do incomplete, duplicate and unexpected inputs stay visible?
Setup: are the steps, module versions and prerequisites easy to follow?
Upkeep: who notices outages and can restore the workflow?
Data flow: which data leaves the system, and which permissions does it need?
Total cost: software, checking, setup and operation, each considered separately.
Export and limits: can readers check the files and repeat the trial?

The criteria come before the result. A commission is not a selection criterion. Two implementations of the same case are not a complete market comparison. There is no star rating here and no blanket test winner.

Three kinds of statement

A trial we ran ourselves: A specific software version processes the documented synthetic inputs. The report and the outputs state the scope that was actually tested. A local run is not a test in the cloud or in sustained operation.
A checked file and documented instructions: Structure, fields and expected values can be traced. For the Make blueprint, the real cloud run is still outstanding. We do not derive any claim of hands-on Make experience from it.
Decision aid: Manual work, cloud and self-hosting are weighed against disclosed criteria. The calculator uses your inputs and supplies neither measured vendor prices nor a forecast of success.

Download the test files and report →

How this section is funded

Dargel Solutions created this test pack itself and provides it free of charge. We take part in the n8n affiliate programme and, since 10 September 2026, also in the Make affiliate programme. Affiliate links to both vendors appear on the overview page of this section and on the n8n and Make guide pages. This means a commission can arise with both tools compared. In the German edition of this track, Make was linked only neutrally until 9 September 2026; that uncompensated counterweight no longer exists, which is why the criteria on this page and the disclosed limits of the evidence are now the only safeguard against a commission incentive. Every link that can earn a commission is labelled “Advertisement / affiliate link” right at the link, together with a note on the possible commission. If a programme ends or is suspended, a neutral vendor link with the note “no commission” replaces the affiliate link.

If you take out a qualifying plan through an affiliate link, Dargel Solutions may receive a commission. This implies neither a price advantage for you nor that the tool suits your case. The link status does not determine our assessment. A manual or uncompensated solution may be the right choice.

Dargel Solutions also sells its own digital products. These are our own offers, not an independent second recommendation. The package you can download here requires no purchase, no Dargel account and no newsletter sign-up.

What happens on this website

The decision tool calculates locally in your browser and does not store its inputs. Downloads are ordinary files. Vendor links lead to third-party websites. The n8n affiliate link first passes through the partner service PartnerStack on its way to the n8n website and contains our partner ID so that a referral can be attributed. The Make affiliate link goes directly, without an intermediate service, to the sign-up form on make.com and carries our partner ID as part of the address. For this section we use no partner pixels, no redirect service of our own, no visitor-specific link IDs and no additional outbound counters.

The website's existing page-view counting remains in place. How it works is described in our privacy policy. On a linked vendor website, the data processing described there applies. An affiliate link does not replace consent for any tracking used there.

When the workflow changes

The download carries a version number and file hashes. Relevant changes to modules, vendor rules or outputs require a new test. We review the software examples and programme information monthly and whenever there is a concrete sign of a change, and the key sources after 90 days at the latest.

If an error changes the recommendation, we mark the affected version or withdraw the affected link. Instructions that are still useful may stay online. Report a discrepancy via the contact page with the file version, the error text and a synthetic example. Do not send credentials or real customer requests.

Editing and responsibility: Dargel Solutions. The texts, this English edition and the test artefacts were created with AI assistance. The evidence available here names the checks that were carried out and the limits that remain open. This English edition is dated 25 September 2026; the German version it translates is dated 25 September 2026.

Documentation for further reading

Public programme descriptions are not individually agreed contract terms. We do not publish a commission calculation here based on durations that have not been clarified.

If you already run an agent workflow, the practical guide How to Audit a Passing Check in an AI Agent Workflow helps you examine one specific result requirement. A worked example and a free worksheet take you from the expected result to a finding you can back up.