OpenAPS (Open Artificial Pancreas System) is an open-source, algorithm-driven system that automates insulin delivery by integrating a continuous glucose monitor (CGM), an insulin pump, and a miniature computer. While the default settings provide a safe starting point, every person’s diabetes is unique. Achieving stable blood glucose, reducing hypoglycemia, and maximizing time in range requires deliberate customization tailored to your daily patterns, sensitivities, and lifestyle. This guide walks you through the key parameters, data analysis techniques, and a systematic method to adapt OpenAPS to your individual needs. By understanding the underlying math and making incremental changes based on real-world results, you can transform the system from a generic algorithm into a personal diabetes assistant that works for you.

Understanding Core OpenAPS Settings

The heart of OpenAPS consists of several interlinked parameters that govern insulin delivery decisions. Before you adjust anything, it’s essential to know what each setting does, how they interact, and what the numbers mean.

Target Glucose Range

The target glucose is the blood sugar level the system aims to maintain. It is defined by a low target and a high target. The low target is the level below which the system will reduce or suspend insulin delivery; the high target is the level above which it will increase basal or deliver micro-boluses. Typical ranges are 100–110 mg/dL (5.6–6.1 mmol/L) for most users, but those with hypoglycemia unawareness or frequent overnight lows may set a higher low target, e.g., 110–130 mg/dL (6.1–7.2 mmol/L). The target can also be temporarily adjusted for exercise, sleep, or high-risk situations using the temp_target feature.

Insulin Sensitivity Factor (ISF)

ISF tells the system how much one unit of insulin lowers your blood glucose. For example, an ISF of 40 means 1 unit drops glucose by 40 mg/dL. This value can vary throughout the day due to circadian rhythms, activity, and hormone cycles. In OpenAPS, ISF is stored as a time‑based set of values (e.g., different settings for morning, afternoon, and night). To find your personal ISF, you can start with the “1800 rule” (dividing 1800 by your total daily insulin dose) and then fine‑tune by observing how correction boluses affect your glucose after meals. A common mistake is setting ISF too low (causing over‑correction) or too high (leaving stubborn highs).

Carbohydrate Ratio (Carb Ratio)

Carb ratio defines how many grams of carbohydrate are covered by one unit of insulin. A starting point might be 10 g/U, but many people need ratios that differ for breakfast, lunch, and dinner due to variations in insulin resistance throughout the day. Inaccurate carb ratios lead to post‑meal spikes (ratio too high – not enough insulin) or early post‑meal lows (ratio too low – too much insulin). Use at least a week of meal data to identify patterns, then adjust in increments of 1–2 g/U.

Basal Rates

Basal insulin provides a constant background infusion. OpenAPS adjusts basal rates within limits you set: a maximum basal rate (the highest temporary rate allowed) and a maximum insulin‑on‑board (IOB) limit to prevent stacking. Your basal profile should reflect your natural insulin needs, especially overnight. To verify your basal, do a “low‑glucose suspension test”: let the system run in low‑basal or suspend mode for a period (e.g., overnight fast) and see if glucose stays flat. If it rises, increase basal; if it falls, decrease. Adjust by no more than 10–15% per change and re‑test after 2–3 days.

Duration of Insulin Action (DIA)

DIA is the time insulin remains active in your body, typically 5–6 hours. Setting DIA too short makes the system think insulin disappears faster than it really does, causing unnecessary stacking. Setting it too long means the system won’t correct high glucose soon enough. OpenAPS uses DIA to calculate IOB and decide when to reduce or increase delivery. Most users find 5–6 hours works well, but you can test by comparing the predicted IOB with actual glucose drops after a large bolus.

Maximum Basal Rate and Maximum IOB

Max basal rate limits how much the system can increase background insulin. A common starting point is 2–3 times your highest scheduled basal rate. Max IOB caps the total amount of active insulin the system is allowed to deliver (excluding your manual meal boluses). Start conservatively (e.g., 2–3 units) and increase gradually as you gain confidence in your settings. These two parameters are your safety net – keep them realistic to avoid dangerous insulin stacking.

Analyzing Your Personal Diabetes Data

Effective customization begins with understanding your current patterns. Without data, adjustments are guesswork. You need to collect information from your CGM, pump, and lifestyle logs.

Nightscout is the primary dashboard for OpenAPS users. It aggregates CGM data and pump history into graphs and reports. Key reports to examine:

  • Daily Glucose Profile – overlay several days to see recurring patterns (e.g., dawn phenomenon, post‑meal spikes).
  • Time in Range histogram – shows percentage of time in low, target, and high zones.
  • Low/High Frequency – identify times of day when you most often go low or high.
  • Deviation Report – compare predicted glucose with actual arrival to see if the system’s assumptions (ISF, carb ratio) are accurate.

Export raw data from Nightscout’s “Reports” section for deeper analysis in a spreadsheet. Learn more about Nightscout reports to uncover subtle trends.

Considering Lifestyle Factors

No algorithm can account for everything. Log exercise sessions, menstrual cycle phases, stress, illness, and alcohol consumption. For example, intense strength training can increase insulin sensitivity for 12–24 hours, while aerobic exercise may cause a delayed drop. If you regularly bike commute, your ISF may need a distinct setting for that period. Also consider your sleep quality – poor sleep often raises overnight insulin resistance. Use a simple app or a notebook to track these factors alongside your CGM data.

Auditing Your Current Settings

Before making changes, export your current preferences file (usually preferences.json on the rig). Compare your basal rates to manual calculations using total daily dose (TDD). Standard formulas can provide a sanity check:

  • 1800 rule for ISF: 1800 ÷ TDD = approximate ISF (mg/dL per unit).
  • 500 rule for carb ratio: 500 ÷ TDD = approximate grams per unit.
  • Basal as percentage of TDD: typically 40–50% of TDD comes from basal. If your basal is much higher or lower, investigate.

These are starting points; your actual settings will differ. Document the current values in a change log.

Step-by-Step Customization Process

Follow this systematic approach to modify your OpenAPS settings. Always test one change at a time and allow 48–72 hours to evaluate each adjustment. Rushing leads to confusion and risk.

1. Backup Your Configuration

Copy your entire rig’s configuration directory. Run openaps report save settings-report.json or download via the web interface. Store the backup off‑device in case you need to revert quickly.

2. Adjust Basal Rates First

Basal rates are the foundation of the system. If your basal profile is wrong, all automated corrections will be built on a shaky base. To test basal accuracy, run OpenAPS in low‑basal mode (or suspend) for an overnight fast. If glucose rises more than 10 mg/dL, your basal is too low; if it falls, too high. Adjust the specific time period by 10–15% and re‑test. Repeat until overnight glucose is flat (±5 mg/dL). Once basal is stable, move on to ISF.

3. Tune Insulin Sensitivity Factor

After basal is correct, evaluate ISF. Look at correction boluses: when your glucose is high and the system delivers a bolus, compare the predicted drop with the actual drop. If the system consistently under‑corrects (glucose stays higher than predicted), your ISF number is too high (i.e., each unit has less effect than assumed); decrease the ISF value. If it over‑corrects, increase the ISF value. Many users need different ISF for different parts of the day. Use Nightscout’s “Deviation” report to see systematic errors.

4. Set Carbohydrate Ratios

Carb ratios are often the most variable. Analyze 5–7 days of meal data. For each meal, record pre‑meal glucose, carbs entered, and the glucose at 2 and 4 hours post‑meal. If you see a consistent rise after breakfast, that ratio is likely too high (not enough insulin). If you go low before the next meal, the ratio is too low. A typical pattern is a higher ratio (more insulin per gram) for breakfast and a lower ratio for dinner. Adjust by 1–2 g/U at a time and observe for a few days.

5. Configure Target Glucose Range

Set your low and high targets realistically. Most adults aim for 100–110 mg/dL. If you have frequent lows (especially overnight) or hypoglycemia unawareness, raise both targets to 110–130 mg/dL. You can also set temporary targets: for exercise, pre‑set 140 mg/dL; for sleep, 120 mg/dL. Use the temp_target profile in Nightscout to automate these changes.

6. Set Maximum Basal Rate and Max IOB

Max basal rate should be 2–3 times your highest scheduled basal rate. Start with 2× to be safe. Max IOB should initially be low – e.g., 2 units – and only increase after you see how the system behaves during high‑correct scenarios. Note: Max IOB includes only insulin delivered by the loop (temp basals and SMB), not manual boluses you give for meals.

7. Use Autosensitivity (Auto‑Sens)

Enable autosens_enabled in your preferences. This feature dynamically adjusts basal, ISF, and target based on recent glucose deviations. It helps compensate for daily fluctuations like illness, stress, or activity changes. Monitor how much auto‑sens adjusts (you can see the autosens ratio in Nightscout). If it consistently leans one direction (e.g., always increasing effectiveness), your underlying settings need revision. Auto‑sens is not a substitute for good base settings – it’s a fine‑tuning layer.

Advanced Customizations for Experienced Users

Once your core settings are well‑tuned, consider these advanced features to further optimize your system. These require a solid understanding of the algorithm and careful testing.

Super Micro Bolus (SMB)

SMB enables the system to deliver small boluses of rapid‑acting insulin every 5 minutes, not just temporary basal adjustments. This can flatten post‑meal peaks and respond quickly to rising glucose. However, it increases risk of insulin stacking. Enable SMB only after your basal, ISF, and carb ratios are very well tuned – typically after several weeks of stable control. Set smb_enabled to true and define min and max limits for SMB size (e.g., 0.05 U minimum, 0.5 U maximum). Monitor closely for the first few days.

Unannounced Meals (UAM)

UAM mode allows the system to handle meals without you entering carbohydrates. It detects glucose rises and delivers micro‑boluses or temporary basal increases to limit spikes. This is useful for unpredictable eating (e.g., children, finger foods). It requires very tight basal settings and a low risk of missed lows because the system may not respond fast enough to fast‑absorbing carbs. Combine with SMB for best results, but test with small meals first.

Temp Targets for Activity and Sleep

Set up multiple temporary target profiles that you can activate via Nightscout or a mobile app. For example:

  • Activity: 140 mg/dL for 2 hours – reduces insulin delivery to prevent lows during exercise.
  • Sleep: 120 mg/dL low target – reduces overnight hypoglycemia risk.
  • Pre‑meal: 90 mg/dL low target – allows more aggressive correction before a meal.

You can automate these with IFTTT or custom scripts that trigger based on calendar events or heart rate data.

Dynamic ISF and Basal from TDD

Some community forks of oref0/oref1 implement dynamic ISF and basal calculations that update based on total daily dose (TDD) from the last 1–3 days. This can help adapt to changing insulin needs without manual adjustments. To use this, you need to modify your rig’s code (Git, Python, and basic programming skills). If you’re comfortable, explore the oref1 branch or custom forks like pHATtrick. Always test with low max IOB first.

Troubleshooting Common Issues

Even with careful customization, problems arise. Here are solutions for frequent challenges:

Persistent Overnight Highs

If you wake up high every morning, check your overnight basal rates. Use Nightscout’s “Daily Profiles” overlay to see if the rise starts before dawn (dawn phenomenon) or after midnight. Adjust the basal segment that precedes the rise. Also consider ISF – if dawn phenomenon is strong, you may need a temporary target of 90 mg/dL before bed or a more aggressive ISF overnight.

Recurring Post‑Meal Lows

Lows within 2–3 hours of eating often indicate a carb ratio that is too low (too much insulin), or a DIA that is too long (insulin still active when glucose drops). Try reducing the meal‑time ratio by 1–2 g/U. Also check if your ISF is too aggressive – if corrections after meals overshoot, increase the ISF number (make each unit less powerful).

System Over‑Correcting for High Glucose

If the system delivers too much insulin when you go high, your ISF is probably too low (the number is too small). Increase ISF (e.g., from 40 to 50) so each unit has less of an effect. Also verify your max IOB – if it’s set too high, the system can stack more insulin than needed. Reduce max IOB temporarily.

Too Many Suspensions (Low Basals)

If you see frequent zero‑temp basals (suspensions), your basal rates may be too high, or your max IOB may be too high for your needs. Lower the basal rate in the affected time block. Also check your temp target – if you’re using a lower target, the system will cut insulin more aggressively.

For more community‑based support and templates, visit the official OpenAPS documentation. Explore the oref0 repository on GitHub for technical details and issue trackers.

When to Consult Your Healthcare Team

OpenAPS is a powerful tool, but it is not a substitute for medical advice. Always inform your endocrinologist or diabetes educator about your customization efforts, especially when changing basal rates, ISF, or enabling SMB/UAM. Bring data visualizations from Nightscout to appointments to support your rationale. If you experience frequent severe lows, Diabetic Ketoacidosis (DKA), or a rising A1c despite adjustments, seek professional guidance immediately. With proper collaboration, your healthcare team can help validate your changes and ensure long‑term safety.

Conclusion

Customizing your OpenAPS settings is an ongoing, data‑driven process that turns a generic algorithm into a personalized diabetes management tool. Start with the fundamentals – basal rates, ISF, carb ratios, and target glucose – and adjust one parameter at a time while observing trends over several days. Use Nightscout and lifestyle logs to identify patterns, then apply incremental changes. Once your core settings are stable, explore advanced features like SMB, UAM, and dynamic adjustments to further improve outcomes. Always keep safety as your top priority: set conservative limits, use autosensitivity, and maintain open communication with your healthcare team. With patience and methodical iteration, your OpenAPS system will adapt to the unique rhythms of your life, giving you more freedom, fewer lows, and better time in range.