Docs & Roadmap — PRESTILON

STEP 0

📋 System Requirements

Before installing PRESTILON, make sure your machine meets these requirements. The installer handles the rest automatically.

🪟

Windows 10 / 11

64-bit · version 1909 or newer

Required

🌐

Local Network Access

Printer must be reachable by IP on LAN

Required

🖨️

Klipper + Moonraker

Running on each printer

Required per printer

💾

~200 MB disk space

For backend + local database

Minimal footprint

💡 Don't know if your printer runs Klipper? Check your printer's web interface — if you see a URL like http://printer.local or a Mainsail / Fluidd dashboard, you're good to go.

STEP 1

⬇ Install PRESTILON

Download and run the one-click Windows installer. It takes care of everything — no terminal commands needed.

Create a free account

Head to prestilon.com/login and sign up. This links your license to the installer — no credit card required for trial.

Download the installer

Go to your Account page → Download tab and grab the latest PRESTILON-Setup.exe. You'll be notified by email when it's available.

⏳ The Windows installer is currently in development. Sign up to get notified when it's ready.

Run the installer

Double-click PRESTILON-Setup.exe and follow the prompts. The installer sets everything up automatically and launches the backend as a Windows service.

Open the dashboard

Once installed, the dashboard opens automatically in your browser at http://localhost:3002. Bookmark it — that's your PRESTILON home.

STEP 2

🔌 Connect Your Printers

Adding a printer takes about 30 seconds. You just need its local IP address and to confirm it's running Moonraker.

Find your printer's IP

Open your printer's web interface (Mainsail or Fluidd) and check the address bar, or look in your router's device list. It usually looks like 192.168.1.xxx.

Add printer in PRESTILON

In the dashboard go to Printers → Add Printer, enter the IP address and a name. PRESTILON auto-detects the Moonraker API and connects via WebSocket.

Watch it connect live

Within seconds you'll see real-time temperatures, printer status, and live telemetry appear on the dashboard. The AI engine starts collecting data immediately.

🔄 PRESTILON maintains a persistent WebSocket connection. If your printer restarts or goes offline, it reconnects automatically — no action needed.

STEP 3

▶ First Run Checklist

After connecting your printer, here's what to do next to get the most out of PRESTILON from day one.

🧠

Let AI warm up

The AI engine needs 1–2 print cycles to establish baseline telemetry. Health scores and anomaly detection become meaningful after that initial calibration period.

🤖

Set your AI autonomy level

In Printer Settings → AI Control choose how much autonomy you want: L0 (observe only), L1 (recommendations), or L2 (auto-correct). Start with L1 if you're unsure.

✈️

Configure Telegram alerts

Go to Settings → Notifications to connect your Telegram account and receive instant alerts for anomalies, errors, and job completions directly in Telegram.

🔬

Run a hardware diagnostic

Open your printer detail page and go to the Diagnostics tab. Run the full suite once to get a baseline health score and catch any existing issues early.

FAQ

💬 Common Questions

Does PRESTILON work with Bambu Lab printers?

Bambu printers use a proprietary protocol and don't run Klipper/Moonraker. PRESTILON currently requires Klipper + Moonraker. Bambu support is on the roadmap.

Can I run PRESTILON on multiple machines?

The backend runs on one Windows machine and connects to all your printers over the local network. Hardware bindings per plan are shown in your Account → Hardware tab.

Does PRESTILON run constantly in the background?

Yes — the backend runs as a Windows service and starts automatically with your PC. It uses minimal resources at idle (< 2% CPU, ~120 MB RAM) and ramps up only when a printer is actively printing.

What happens if my internet goes down?

Everything keeps working. The AI engine, dashboard, telemetry, and anomaly detection are all fully local. Internet is only needed for license validation (done once at startup) and sending telegram alerts.

Still stuck?

Post in the Community forum or send us a message via Support. We respond within 24 hours on business days.

AI ENGINE

🧠 How the AI Works

The entire AI stack runs on your Windows machine — zero cloud inference latency. It processes five live telemetry streams continuously and makes decisions in under 1ms.

📈

CUSUM Control Charts

Cumulative Sum statistical process control detects small, sustained shifts in printer behavior — catching belt wear, clog onset, and thermal drift at the earliest statistical moment, long before visible failure.

🌊

Kalman Filtering

Noise-filters raw sensor data in real time to produce clean, accurate telemetry estimates. This makes the anomaly detection dramatically more sensitive without producing false positives from normal sensor noise.

🏆

Health Score (0–100)

A per-printer composite score synthesizing anomaly severity, diagnostic results, wear model estimates, and maintenance history. Updated after every print cycle. The score is comparable across your entire fleet.

⚠️

Failure Prediction — 5 Modes

Separate prediction models for: thermal runaway, belt degradation, clog risk, bed adhesion failure, and extruder wear. Each outputs a risk probability and estimated time to failure so you can act before a print is ruined.

MONITORING

📡 Live Telemetry

PRESTILON maintains a persistent WebSocket connection to every Moonraker instance. Data arrives in under 50ms and is stored locally in a high-speed time-series database.

🌡️

Temperature Tracking

Real-time hotend, bed, and chamber temperatures with historical charts. Tracks both actual and target values to detect thermal runaway and heating inconsistencies.

📊

Print Progress & History

Live print progress with filename, layer count, time remaining, and estimated completion. Full job history per printer with success/fail tracking and statistics.

💨

Fan & Movement Data

Part cooling, hotend, and controller fan speeds. Live XYZ position and movement velocity — used by the AI to detect belt slack and axis binding patterns.

AUTONOMY LEVELS

🤖 3-Level AI Autonomy

You decide how much the AI is allowed to do. Set a level per printer — keep some under observation, let others self-correct.

👁

L0 — Observe

The AI monitors, scores, and records everything but takes no action and makes no suggestions. Pure data collection mode — useful while you're learning the system.

💡

L1 — Recommend

The AI surfaces anomalies, predictions, and suggested corrective actions as notifications on the dashboard and by email. You stay in full control — the AI advises, you decide.

⚡

L2 — Auto-Correct

The AI autonomously executes safe corrective G-code actions — fan speed adjustments, print pauses, emergency stops — without waiting for approval. All actions are logged with full reasoning.

DIAGNOSTICS

🔬 Hardware Diagnostics — 6 Tests

Each diagnostic sends real G-code to your printer and benchmarks the result against a fleet baseline. Run the full suite when setting up or after any hardware change.

📡

Firmware Status Check

Verifies Klipper firmware version, MCU health, and Moonraker API response times.

🌡️

Thermal Stability Test

Heats hotend and bed to target temps and measures PID stability, overshoot, and time-to-temperature against baseline.

📐

Axis Movement Test

Runs calibrated XYZ moves and analyses response curves for belt tension, backlash, and binding.

🧵

Extrusion Flow Analysis

Extrudes a measured filament length and cross-references with extruder step counts to detect slippage and clog risk.

🟦

Bed Level Verification

Reads bed mesh data and scores mesh variance, identifying high spots and unlevel sections before a print starts.

💨

Fan Response Test

Ramps part cooling and hotend fans through their speed range and checks for stalls, noise anomalies, and RPM deviation.

PUBLIC ROADMAP

✓ Complete

Phase 1 — Core AI Engine

✓

CUSUM anomaly detection

Statistical process control across 5 telemetry streams

✓

Kalman filter telemetry pipeline

Real-time noise reduction on sensor data

✓

Failure prediction models (5 modes)

Thermal runaway, belt, clog, adhesion, extruder wear

✓

AI Health Score (0–100)

Composite per-printer quality score

✓

WebSocket Moonraker integration

Persistent real-time connections to all printers

⚡ In Progress

Phase 2 — Dashboard & Installer

✓

Fleet dashboard (web)

Real-time multi-printer overview with live telemetry

✓

Bed mesh visualization

2D heatmap + history of mesh calibrations

✓

AI control panel (L0 / L1 / L2)

Per-printer autonomy level configuration

✓

Maintenance prediction & logging

Component wear estimates + manual change registration

⟳

Windows one-click installer (.exe)

Auto-installs backend as a Windows service — no CLI required

TBA

⟳

Hardware diagnostics suite (6 tests)

Full automated benchmark against fleet baseline

TBA

↑ Up Next

Phase 3 — Polish & Ecosystem

○

Webcam live view integration

In-dashboard camera feed per printer with AI frame analysis

TBA

○

Mobile app (iOS & Android)

Fleet overview, alerts, and remote pause/stop from your phone

TBA

○

Cost & filament tracker

Track filament usage, cost per print, and fleet ROI

TBA

○

Print scheduling & queue

Queue jobs across multiple printers from one interface

TBA

◌ Future

Phase 4 — Expanded Hardware Support

○

Bambu Lab printer support

Native integration via Bambu's LAN-mode API

TBA

○

OrcaSlicer / PrusaSlicer integration

Import slicer profiles for smarter AI baseline calibration

TBA

○

macOS / Linux backend

Cross-platform support beyond Windows 10/11

TBA

○

Fleet API & webhooks

Programmatic access to fleet data for custom integrations

TBA

COMMUNITY VOTE

What should we build next?

Vote on the features you want most. The top-voted items get moved up the roadmap. Results update in real time.

📷 Webcam AI analysis

Detect spaghetti, blobs, and layer shifts from the camera feed automatically

72% · 148 votes✓ Voted

📱 Mobile app

iOS & Android app with fleet overview, push alerts, and remote pause/stop

65% · 133 votes✓ Voted

🐼 Bambu Lab support

Native integration for Bambu printers via LAN-mode API

58% · 119 votes✓ Voted

🍎 macOS / Linux support

Run the PRESTILON backend on Mac or Linux instead of Windows only

49% · 100 votes✓ Voted

📋 Print scheduling & queue

Queue and schedule jobs across multiple printers from one place

41% · 84 votes✓ Voted

🔗 Fleet API & webhooks

Programmatic access to fleet data for custom integrations and automation

34% · 70 votes✓ Voted

Discuss on Community →

Everything you need to get started.

What should we build next?