Skip to content
FixNook
Sign inStart free
The full pipeline

Every layer exists to answer: what do I replace, and how sure am I?

Parser, rule engine, knowledge base, and AI work in sequence — each stage grounding the next — so the final repair plan is fast, explainable, and honest about its confidence.

High-speed Go parser

Panic Full logs are dense, inconsistent, and vary across chip generations. The parser is built in Go for one job: extract every useful signal, fast.

Parses hundreds of Apple panic format variants across SoC generations.
Extracts 40+ fields: panic string, bug type, kernel and build versions, device model, boot stage, chip revision, socId, baseband, watchdog and more.
Hardware domains covered: USB / Lightning / USB-C, charging, PMIC, SMC, thermal, battery, audio, camera, display, touch, Wi-Fi, Bluetooth, Face ID / TrueDepth, NAND / storage, power rails, sensors, I²C, SPI, GPIO, temperature, voltage, and current.
Structured output feeds the rule engine and AI — no regex guessing downstream.
At a glance
GoMillisecond parse times40+ fieldsHundreds of formats

Deterministic rule engine

Before any AI runs, curated repair rules match against the parsed signals. Rules are consistent, testable, and always explain why they fired.

USB + overcurrent + charging + lightning patterns point to the charging port, charging IC, or USB controller — and say so explicitly.
Thermal + battery + temperature patterns point to the battery, battery flex, thermistor, or power IC.
Every match includes its reasoning, so you can audit the suggestion instead of trusting a black box.
Deterministic rules remain the primary source of consistency as the platform's machine-learning feedback loop matures.
At a glance
Explainable matchesRuns before AIAuditable

Repair knowledge base

A structured repair database, not a forum scrape. Each record captures what a senior tech would write down.

40+ repair records covering the most common panic signatures.
Per record: issue description, common symptoms, related panic codes, possible causes, repair procedure, testing procedure, required parts, difficulty, estimated repair time, warnings, technician notes, and references.
Matched records are injected into the AI's context so suggestions stay grounded in real repair practice.
At a glance
40+ recordsParts & proceduresDifficulty + time estimates

AI analysis that shows its work

The AI receives the parsed data, rule matches, knowledge-base hits, and the original log — then explains the panic in plain English.

Returns a summary, confidence score, likely faults, affected components, recommended tests, recommended repairs, difficulty, estimated time, warnings, and possible alternative causes.
Cites evidence from the log for every claim and states uncertainty instead of guessing.
Never invents a fault: low-confidence results say what to check next rather than naming a part.
At a glance
Confidence scoreCited evidenceAlternative causes

Reports for the bench and the customer

Turn an analysis into paperwork your customer can understand and your bench can act on.

Generate PDF, HTML, and JSON reports from any analysis.
Shareable URL for remote review — with a configurable expiry.
Two variants: a printable customer report and a detailed technician report.
At a glance
PDFHTMLJSONShare links

API-first platform

Everything the dashboard does, the API does. One public REST API, documented with OpenAPI.

Secure API keys generated from your dashboard, with rate limiting built in.
Swagger / OpenAPI documentation published alongside the API.
Designed for future Electron / Tauri desktop apps, Flutter / React Native mobile apps, and third-party integrations.
At a glance
REST + OpenAPIAPI keysRate limited

Security and privacy by default

Logs can contain device identifiers — the platform treats them as sensitive customer data end to end.

Virus scanning, file validation, duplicate detection, and signed upload URLs on every upload.
Encrypted storage, audit logs, and role-based access control across the platform.
Diagnostic assistance only: no security bypass, no unlocking, no modification of Apple security mechanisms — enforced in the rule engine, AI layer, and terms of service.
Feedback used to improve suggestions is anonymized before storage.
At a glance
Encrypted storageAnonymized feedbackDiagnostic-only

Stop guessing. Start with the log.

Your first 10 analyses every month are free — no card required. Upload a panic log and see the full pipeline: parse, rules, AI, repair plan.

Start freeRead the getting-started guide