Founder runbook
Architecture, setup, deployment, maintenance and safe-change guidance for the person responsible for the product.
Free foundationA living blueprint for the product you shipped
You built the app. Now understand it, improve it and teach people to use it.
Explain My Build studies your code, build history and running product to create a founder runbook, a user guide and a verified build-health record.
Free public-repository preview. No payment card. You choose the analysis path and see where data goes before a run starts.
PROJECT / PAPERTRAILanalysed
Application identity
Client request workspace
Static structure, build history and runtime observation agree on 47 claims. A production audit is still missing.
NEXTRun the scoped build-health auditInspect example evidence →
codehistoryruntimeaudit pending
The founder's problem
The code runs. The operating knowledge is scattered.
AI can help you ship before you have time to learn every route, dependency and deployment decision. The result may work while still being difficult to operate, explain or change safely.
Explain My Build turns the evidence already inside the product into a reviewable working record—and labels the gaps it cannot support.
Four connected deliverables
One study of the app. Four ways to use what it found.
User guide
Outcome-based tasks and onboarding written for each authorised persona, with owner review before publication.
Review requiredVerified improvements
A ranked list tied to observed code, runtime or audit evidence—not a generic collection of best practices.
Evidence linkedBuild health
A versioned record of audit coverage, regressions, unresolved unknowns and what changed for users.
Living recordEvidence before prose
Claims gain strength when independent sources agree.
The system preserves Docent, Walker and Auger as separate tools with versioned contracts. Explain My Build coordinates their outputs; it does not blur them into one unexplained score.
CODEStatic structureroutes · models · config
HISTORYBuild intentcommits · churn · decisions
WALKERRuntime observationpersonas · journeys · failures
AUGERBuild-health auditverified · manual · unknown
VERSIONED APP MODELClaim + source + reason
Choose the privacy path first
“Where your data goes” changes with the path you choose.
Every route should be understood before analysis. These summaries describe the designed data flow; the run confirmation shows the selected AI provider and retention details.
Public GitHubHosted
The repository is temporarily cloned to Explain My Build infrastructure, analysed, and deleted after the run. Derived evidence follows the AI-provider path shown at confirmation.
Full data path →Private GitHubRead-only App
You select repositories and see the requested permissions. Encrypted Git data is relayed transiently through Explain My Build and processed in browser memory; it is not intentionally retained.
Full data path →Local folder / ZIPBrowser memory
Source files are read in browser memory and are not uploaded to Explain My Build. The confirmation identifies any remote AI provider that will receive a derived evidence digest.
Full data path →CLI / desktopStrongest control
Analysis runs on your machine. A configured remote provider may receive a derived digest; a local model keeps that provider step on-device. Desktop release remains gated on signing and notarisation.
Full data path →A guide people can use
Tasks, prerequisites and outcomes—not a Markdown dump.
Each persona gets only the journeys supported for that role. The owner can inspect evidence, edit the draft and control publication.
Open the interactive example →Papertrail helpRequesterReviewerAdmin
TASK 04
Send a request for review
Before you start: your workspace must have at least one reviewer.
- Open the request you want reviewed.
- Choose Send for review.
- Confirm the assigned reviewer and due date.
Expected result: the request moves to “In review” and the reviewer is notified.
Verify in your app: notification delivery was not observed in the last runtime walk.
BUILD HEALTH / 6 RUNS2 regressions
Changed Admin export route now returns 500
Unknown Payment webhook replay not assessed
Improved Keyboard focus restored on modal close
Watch the product change
A clean result only describes checks that actually ran.
Build health separates automatically verified findings, items requiring manual confirmation and dimensions the audit could not assess. Tool failure is never converted into a pass.
Three specialists, one coordinated run
Docent explains. Walker observes. Auger challenges.
Discovers the application surface and build intent, then constructs the final versioned AppModel.
Observes authorised routes and personas, emitting a safe manifest plus protected sensitive evidence.
Runs available checks, deduplicates findings and reports verified, manual and cannot-assess partitions.
Accuracy and limitations
An evidence-backed draft, not an oracle.
Claims retain their source, coverage and confidence. Missing runtime access, failed audit tools and business intent that exists only in a founder's head remain explicit unknowns.
Explain My Build does not replace legal, security or compliance review. A “clean” audit means the checks that completed did not find a matching issue; it is not a guarantee of safety.
Read the complete limitations →Plans
Understand your own app for free. Commercial publication is paused.
Free
$0 no card
- Founder runbook
- Evidence summary
- Limited verified improvements
- Public-repository preview
- One project snapshot
Guide Pass
$29 proposed once
- One project deliverable
- Seven-day revision window
- Permanent access to generated exports
Pro
$19 proposed / month
- Living user guide
- Publication and share controls
- Build-health history
- Repository management
No paid plan is currently offered or charged. Prices describe the product direction and may change before reviewed commercial terms are published.
Questions
Before you analyse a build
Does every claim come from code?
No. Claims can be supported by static code, build history, authorised runtime observation or audit evidence. The workspace shows the source and flags claims that remain weak or unsupported.
Does private GitHub data reach Explain My Build?
The current browser path relays encrypted Git traffic transiently through a same-origin proxy. It is not described as zero transit. Choose a local folder, ZIP or CLI when source must not transit Explain My Build infrastructure.
Is this a penetration test?
No. Auger coordinates available static, dependency, configuration and authorised dynamic checks, but incomplete coverage and unavailable tools are reported as cannot-assess. High-stakes systems still need qualified human review.
Can I buy the user-guide product today?
Not yet. Paid activation is disabled while independent legal terms and privacy disclosures are completed. The free public-repository preview remains available without a card.
Start with what can be observed