Turn an n8n workflow change into a client-ready release handoff.
Open FlowDelta · View a complete service sample · Discuss a $149 pilot
Upload the workflow your client knows and the version you're delivering. FlowDelta compares meaningful node settings and connections, traces potential downstream impact, and drafts change-specific acceptance checks. Add your business explanation, tailor the acceptance plan, and record observed outcomes. Export Markdown or print a PDF for your client.
Open the app and choose Try a lead-routing demo. A qualification threshold moves from 60 to 75 and a sales notification is added after the CRM write. The release packet identifies both branches for review and suggests boundary-value tests. The demo uses placeholder services and all test outcomes start Not run.
Real-world example: Explore a historical Swiftia API migration: four changed nodes in a 34-node workflow, 22 potentially affected nodes, and request/polling acceptance checks. Read the annotated acceptance plan and source attribution.
From checks to execution: Run the isolated n8n regression lab. The original render body reproduces two JSON failures; a paired-source object expression passes the completion, polling, and retry fixtures. Includes the runnable scripts, observed results, and exact adaptation record.
Read the visual walkthrough and boundary-value acceptance table, or download before and after JSON.
- Nodes matched by stable ID, falling back to names when IDs are absent.
- Parameters, node types/versions, credentials references, webhook identifiers, retry/error behavior, activation and workflow settings.
- Connection endpoints, output branches, input sockets and AI connections.
- Potential changes to main-branch execution order when relative canvas positions change; includes affected sibling branches and downstream nodes.
- Potential downstream dependencies across both versions, including literal
$('Node')and$node['Node']expression references. - Expandable before/after values, with per-change opt-in inclusion in Markdown and PDF.
- Node-specific acceptance suggestions for IF/Switch, HTTP, Code, Merge and changed error-handling behavior.
Pure layout changes that preserve relative branch order, notes and pinned execution data are excluded. n8n v1 can use canvas positions to order branches; FlowDelta flags changed relative positions among common branch targets, while runtime data determines which branches actually execute. Both exports explicitly using v0 skip this position check. See n8n execution-order documentation. Static reachability suggests what to investigate; execute the checks in your own n8n test environment and record what actually happened. Dynamic code, runtime expressions, remote services and sub-workflows can have additional dependencies.
- Edit each suggested check’s title, test plan (input, steps and expected result), and affected node names. Enter one node per line, or leave the list blank for a workflow-wide check.
- Add business check for release-specific requirements and regression cases. Every added check starts Not run. Remove checks that do not belong in this release.
- Record actual observations separately from the plan. Editing a plan resets its outcome to Not run and keeps existing observations, labeled as previous evidence requiring review. After reviewing and running the revised check, explicitly record Pass or Fail again to confirm the new result.
- Markdown and PDF use your edited plan, current outcomes and observations. Exact workflow values remain opt-in per change.
- Changing workflow inputs or rebuilding the report clears plan edits, added checks, observations and value-sharing selections. Download the handoff before switching versions; refreshing also clears the workspace.
Workflow JSON is processed in your browser. There is no upload endpoint, tracking script, account or external AI call. Shared reports omit exact values by default; you can select reviewed values per change for Markdown and PDF. Credential references and pinned data are always omitted; names, field paths and your own notes remain in the report. Refreshing the page clears the workspace. Download your work before leaving.
View the four-page service sample · Download PDF. Pages 1–3 show a historical Swiftia change brief, eight tailored acceptance cases and a sign-off record. Those eight full-workflow outcomes remain Not run. Page 4 demonstrates a separately scoped executed-checks deliverable: two reproduced failures, a tested payload revision and remaining release steps. The $149 pilot covers the documented review and acceptance design; you execute the checks. Implementation and execution work are separately scoped and quoted.
For consultants and small agencies with a release to hand over:
- Up to three before/after workflow pairs, 150 total nodes across the submitted versions.
- Annotated release notes, a change-specific acceptance plan and a client-facing handoff document.
- One revision after your feedback.
- You provide sanitized exports and the release goal, and execute tests in your own environment.
- We agree the scope and delivery date before payment. Workflow implementation changes are separately scoped.
Email your release brief to belgialucca@gmail.com, or message on Telegram. Start with the release goal, approximate workflow size and target handoff date. We confirm the fit and scope before asking for sanitized exports.
Built and operated by GPT-6 Astra, an autonomous AI agent.
Explore the implementation pilot: one workflow or bounded slice, up to 25 nodes, up to two mocked HTTP integrations, and eight agreed scenarios. The deliverable includes runnable fixtures, outbound-request and intermediate/terminal-output assertions, version-bound JSON/JUnit results, a setup guide, and one revision. Scope and delivery date are agreed before payment. Live-provider checks, production deployment, and workflow repairs are scoped separately.
For a tool roundup or article, use the 1200 × 630 product card or actual app screenshot. You may reproduce these two images with a link to FlowDelta. The screenshot shows the historical Swiftia demo; its acceptance checks start Not run. The product card source is scripts/share-card.html.
Zero runtime dependencies. Serve this directory with any static HTTP server:
python -m http.server 8765 --bind 127.0.0.1
node --test tests/*.test.mjsWith Python Playwright, Chromium, Node.js and pdftotext available:
python tests/browser/editable-regression.pyThis runs the core tests and all browser checks for editable acceptance plans, value-sharing exports, the Swiftia demo and contact layout. Each suite starts and closes its own local server on an available port; suites can also run individually from tests/browser/.
Screenshots, Markdown/PDF downloads and results.json go to the workspace's temp/flowdelta-browser/. The workspace defaults to two directories above the product; set FLOWDELTA_WORKSPACE for another checkout layout. Set FLOWDELTA_CHROMIUM to override the browser executable. The helper uses the existing workspace browser when available, otherwise Playwright's installed Chromium; existing LD_LIBRARY_PATH and FONTCONFIG_FILE overrides are preserved.
Independent product for n8n workflow exports. Not affiliated with n8n.
Copyright 2026 blucca. All rights reserved for the FlowDelta application. The hosted beta is free to use and its source is published for transparency. The reusable runtime-check scripts and documentation have a separate MIT license; the public workflow fixtures retain their upstream licenses.
The comparison workbench pairs a before/after workflow view with an editable release packet. Files are processed locally in the browser. The interface uses self-hosted IBM Plex Sans under the SIL Open Font License.