A substance dose tracker for iOS. Logs what you took, when, and at what dose; overlays pharmacokinetic curves so you can see what's still active; warns about drug interactions before you stack them.

Built primarily for harm reduction: medication adherence, recreational session tracking, and stack safety. Substance data is sourced from TripSit, PsychonautWiki, and the FDA's DailyMed — merged and deduplicated into a single library of ~1,100 substances with dose ranges, routes, durations, and half-lives.

• ~1,100 substances with dose ranges (threshold / light / common / strong / heavy), administration routes, duration profiles, and half-lives.
• Pharmacokinetic timeline. A one-compartment oral PK model estimates absorption rate from each route's onset and renders the resulting concentration curve; doses scale the curve height by amount-to-heavy ratio.
• Interaction warnings. 59 class-based interaction rules covering the common danger pairs (MAOI + stimulant, opioid + benzo, lithium + NSAID, etc.). Warnings appear inline during logging and as a danger window on the timeline.
• Live Activity. Active doses appear on the Lock Screen and in the Dynamic Island with their remaining duration.
• Harm-reduction notifications. Optional ramp-down reminders grouped into 6-hour sessions, so back-to-back doses don't fragment into separate alert chains.
• Daily medication tracking. Prescription / supplement adherence with weekday-aware reminders, separate from the recreational journal.
• Insights. Activity heatmap, usage stats, half-life calculator, per-substance trend charts.
• Custom substances. Add anything not in the merged library — name, category, dose ranges, route, half-life — and it participates in search, interactions, and PK.
• Export. PsyLog-compatible JSON and PDF reports with PK charts.

• iOS 26 or later. The UI is built around Liquid Glass; older iOS versions are not supported.
• Xcode 26 or later with Swift 6.
• Apple developer account for code signing (TestFlight only; not on the App Store).

git clone https://github.com/kageroumado/piru.git
cd piru
open Piru.xcodeproj

Select the Piru scheme and Run. Code signing is configured for the original developer team; set your own DEVELOPMENT_TEAM and bundle identifiers in Xcode's signing settings before building to a device.

For a headless compile check:

xcodebuild -scheme Piru \
  -destination 'platform=iOS Simulator,name=iPhone 17 Pro' \
  build

To run the test suite (Apple Testing framework):

xcodebuild -scheme Piru \
  -destination 'platform=iOS Simulator,name=iPhone 17 Pro' \
  test

There is also a Swift Package Manager CLI in Tools/SubstanceValidator that cross-checks the merged substance library against the upstream APIs; useful when refreshing the cache.

cd Tools/SubstanceValidator
swift run SubstanceValidator validate

Piru/
├── Models/                 Substance, DoseRange, DurationProfile, units
├── Views/                  SwiftUI tabs: Journal, Library, Tools, Insights
├── Data/                   SubstanceLibrary, HalfLifeDatabase, Interactions
├── Services/               TripSit, PsychonautWiki, DailyMed API clients
└── Utilities/              PKModel, RampDownScheduler, exporters

Shared/                     Code shared with widgets + Live Activity
PiruLiveActivityExtension/  Lock Screen / Dynamic Island widget
PiruWidget/                 Home Screen widgets
PiruTests/                  Swift Testing suites
Tools/SubstanceValidator/   CLI for validating substance data

• SwiftUI + SwiftData. @Model types are used directly in views; there is no separate ViewModel layer. Managers (SubstanceLibrary, LiveActivityManager, etc.) are @Observable @MainActor singletons.
• Swift 6 strict concurrency with @MainActor default isolation.
• Substance data lifecycle. Three API clients fetch in parallel, results are merged with a name-normalising deduplicator, and the combined library is cached to substances_cache.json with a 7-day TTL.
• Search. Ranked cascade — exact match → alias → prefix → substring → fuzzy (Levenshtein) — so common names beat obscure ones even when both match.
• PK model. One-compartment oral with route-aware absorption. ka is estimated from each route's onset time; Cmax and Tmax are derived from the resulting curve and overlaid on the timeline.

Issues and PRs welcome, with the caveat above about maintenance pace. Substance data corrections are especially useful — open an issue with the substance name and the source you're using.

If you ship a fork, please don't call it "Piru" — the name and the pink flame icon are pharmacykitty's. Pick your own.

Piru is free software, licensed under the GNU General Public License, version 3 — see LICENSE.

Originally built by pharmacykitty; development is now continued by kageroumado. Pharmacology data courtesy of TripSit, PsychonautWiki, and the NIH's DailyMed. Half-life values are drawn from a hand-curated database of ~1,100 entries; see Piru/Data/HalfLifeDatabase.swift.

Piru is not medical advice. It will not stop you from making a bad decision. Get a test kit, dose low, and have a sober friend.