Personal Relationship Manager (PRM)

A CLI tool for tracking relationships, interactions, and reminders. Privacy-first: all data stays on your machine. No external APIs.

Try it (no install required)

Grab the prebuilt binary for your OS from the latest release, unzip it, and follow the "READ ME FIRST" file inside — no Rust, no Ollama, no build step. This build skips the AI features (see below) so there's nothing else to set up.

Setup (build from source)

1. Install Rust (stable, 1.93+)
2. Install Ollama for AI features, then pull the default model:

   ollama pull llama3.2:3b
   

3. Clone and run:

   git clone https://github.com/petros-p/prm.git
   cd prm
   cargo run
   

Data is stored in .data/prm.db (SQLite). AI features (ai-log, voice-log, inbox) are included by default; build with --no-default-features to skip them (e.g. if you don't want to compile whisper-rs, or Ollama isn't available).

Commands

People

Quick Edits

Circles

Labels

Interactions & Reminders

Other

AI Features

All AI runs locally — no API keys, no data leaves your machine.

• ai-log — Describe an interaction in plain text; Ollama parses it into structured data for review and save.
• voice-log — Record a .wav file; Whisper transcribes it locally, then Ollama parses it.
• inbox — Batch-process free-text notes dropped into a watched folder, through the same parse/review/save flow as ai-log.

Capturing notes from anywhere

PRM watches a local folder — .data/inbox by default, or wherever PRM_INBOX points — for plain .txt/.md files. Each time you run inbox, every file in that folder is parsed by the local AI and offered up for the same review/edit/save flow as ai-log, then moved into inbox/processed (or inbox/failed if it couldn't be read).

Getting a note from your phone into that folder is up to whatever sync tool you already use — PRM only reads a plain folder, so this works the same on any phone or computer:

• iCloud Drive / Google Drive / Dropbox / Syncthing — point the app at .data/inbox and share/save a note into that synced folder from your phone (e.g. the Share Sheet's "Save to Files" on iOS, or "Save to Drive" on Android).
• AirDrop / USB / manual copy — just drop a .txt file in directly.
• Voice memo — transcribe it (or use voice-log directly if the .wav is already on the machine running PRM) and save the transcript as a .txt file in the folder.

There's no server, no port, and nothing that needs to run continuously — PRM only checks the folder when you run it, and the file transport is entirely handled by tools you already have.

Voice log setup

Download the Whisper model (~148MB) and place it at .data/models/ggml-base.en.bin:

mkdir -p .data/models
curl -L -o .data/models/ggml-base.en.bin \
  https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-base.en.bin

Environment variables

Building

cargo build --release    # Optimized binary at target/release/prm
cargo test               # Run all tests (99 tests)
cargo check              # Fast compile check