BEGINNER-FRIENDLY · FREE REFERENCE SAMPLE
Send one anonymized task, workflow, or connected workflow family. Builderlog maps the intended result, the highest-risk failure, three priority fixes, and one acceptance test in plain language. Custom implementation is not currently sold.
CHOOSE THE SMALLEST SAFE STEP
SELF-SERVE · SOLO
Use the workbooks, sixteen-case acceptance matrix, bounded agent-loop receipts, document handoff receipts, empty-branch receipts, ambiguous-retry reconciliation, replay receipts, and verified reference pack to adapt your own workflow. Instant download; no private process details required.
Get the Solo kit →PLAIN-LANGUAGE REFERENCE · FREE
Inspect a synthetic task or workflow. See three prioritized fixes and one acceptance test, then decide what requires human review.
Inspect the free reference sample →Start with $19 if you can work through the templates yourself. View the free reference sample when you still do not know what to fix first. Neither option requires credentials or customer data.
GOOD FIT
✓ One primary outcome and a documented input
✓ Up to three connected n8n workflows
✓ Up to three external services
✓ Anonymous examples that cover valid, invalid, duplicate, and spoofed-approval inputs
✓ Credentials connected by the buyer after delivery
DELIVERY PACKET
An allowlisted input schema, configuration boundary, primary outcome, and fields that public callers are never allowed to control.
Up to three connected n8n workflow files for the confirmed process.
A real state adapter and duplicate receipt—not an idempotency key that is calculated and then forgotten.
Approval fields from public intake are stripped. Trusted reviewer alias, channel, and timestamp persist into the final receipt.
At least seven valid, invalid, replay, state-failure, trusted-review, and authority-spoof cases with observable expected results.
A buyer-runnable verification command, artifact hashes, import steps, state-store choice, recovery actions, and topology notes.
CURRENT N8N EDGE
Portable: Data Table replay ledger for light/moderate, low-concurrency work.
Concurrent: database unique key and atomic insert/upsert when identical events can race.
Scaled: queue-mode and external task-runner readiness notes when the buyer already runs that topology.
AI-assisted: evaluation datasets and trusted human-review timestamps where the workflow uses AI decisions or tools.
MULTI-CLIENT CONTRACT MODE
Shared logic: up to three connected workflow files for one operational process.
Tenant configuration: an allowlisted schema that rejects missing or unknown client settings.
Synthetic canary: one anonymous tenant fixture proving configuration and record isolation.
Rollout receipt: version, tenant alias, outcome, artifact hash, and rollback decision format.
Boundary: no credentials, live customer records, or production deployment across the client fleet.
HARD BOUNDARY
Builderlog does not collect passwords, API keys, private customer data, or production account access. The review excludes workflow implementation, contact scraping, unsolicited outreach, Terms-of-Service bypasses, production migrations, ongoing monitoring, custom front-end applications, and open-ended consulting.
CLOCK + REFUND
Scope mismatch: the paid review is no longer available.
Reference: read the synthetic sample to inspect the intended outcome and acceptance-test format.
Privacy: no intake, credentials, or customer data is requested.
Boundary: the sample is educational reference material.
EVIDENCE BEFORE PAYMENT
Download Builderlog Reliability Contract Pack v2.3: three workflows accepted by the n8n 2.31.0 beta-line CLI importer, an atomic PostgreSQL replay reference, two native evaluation fixtures, and eleven workflow/topology cases. The zero-dependency verifier currently passes 41/41 checks and 11/11 cases. npm's stable line was 2.30.7 at the latest check, and stable-line import acceptance is not claimed. It is a reference—not a paid customer case study.
Inspect the exact delivery shape →READ THE BUILD LOGS
These are working notes from the same reliability system: how the price boundary is set, how a green run can still update the wrong record, and how scheduled output is verified in public.
REDACTED SCOPE INTAKE
This fit check is free and happens before payment. Do not submit passwords, tokens, customer data, private URLs, or production exports. A public messenger handle and an anonymous process description are enough for the first scope decision.
FAQ
One operational outcome described in plain language, one anonymized workflow, or a connected workflow family. It cannot be three unrelated projects.
No. Builderlog uses redacted exports, screenshots, or anonymous samples. Never send credentials or customer data.
No. The paid review is no longer available; the free reference sample remains available for reading.
Three prioritized fixes, one acceptance test, and a written decision about whether to keep the work manual, use the self-serve kit, improve the current workflow, or define a separate implementation scope.
Yes, when they serve one operational outcome. The report maps the boundary; it does not include workflow builds or deployment across client accounts.
Contact scraping, unsolicited outreach, Terms-of-Service bypasses, production migrations, ongoing account management, custom front-end apps, and open-ended consulting.
$0 FIT CHECK · $19 SELF-SERVE · FREE REFERENCE
