# Arthur-Ficial/apfel

The free AI already on your Mac. CLI tool, OpenAI-compatible server, and interactive chat — all on-device via Apple Intelligence. No API keys, no cloud, no downloads.

Repository: https://github.com/Arthur-Ficial/apfel
Canonical: https://ross.abutalabs.com/products/apfel
Homepage: https://apfel.franzai.com
Language: Swift
License: MIT
License Family: permissive
Topics: apple-intelligence, apple-silicon, cli, foundationmodels, llm, macos, macos-26, on-device, swift, homebrew, openai-compatible, tool-calling, unix
Last push: 2026-08-05T14:44:49+00:00

## Health v2 (maintenance only)
Score: 75/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 96, release rhythm 84, longevity 11
- inputs: {"age_days": 162, "days_push": 28, "days_rel": 28, "gap_med": 0.0, "n_releases_24m": 81}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6325, forks 243 (observed 2026-08-28T04:09:41.745356+00:00)

## What it is
apfel is a Swift CLI tool that exposes Apple's built-in on-device Foundation Models LLM on Apple Silicon Macs as a UNIX-friendly command, an interactive chat REPL, and a local OpenAI-compatible HTTP server. It requires no API keys, no model downloads, and no cloud, running entirely on-device on macOS 26+.

## Use cases
- run LLM prompts from the terminal on a Mac without API keys
- pipe text through a local LLM in shell scripts
- serve an OpenAI-compatible local endpoint for OpenAI SDKs
- summarize or translate files on-device for privacy
- get shell commands from natural language with --code
- use tool calling with a free on-device model
- add AI to scripts without cloud costs or data leaving the machine

## When to choose
- you're on an Apple Silicon Mac with macOS 26+ and Apple Intelligence enabled
- you want zero-config, zero-cost, fully private on-device LLM access
- you need a drop-in OpenAI-compatible local backend for existing SDKs
- you want a composable UNIX tool with JSON output, streaming, and exit codes

## When to avoid
- you're on Linux, Windows, or Intel Macs
- you need large context windows beyond 4k-8k tokens
- you need frontier-model quality or multimodal generation beyond basic image understanding
- you need a model you can choose, fine-tune, or swap

## Facets
- artifact type: cli-tool
- maturity: active
- function: llm-inference, cli, http-server, chatbot, agent-framework
- domain: large-language-models, artificial-intelligence, developer-tools, apple-ecosystem
- platform: cli
- tags: apple-intelligence, foundation-models, on-device-llm, openai-compatible, apple-silicon, unix-tool, tool-calling, homebrew, command-line, macos, swift

## Member repositories
- Arthur-Ficial/apfel (main) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:41.745356+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:46:10.425519+00:00, confidence not recorded.
  - readme: https://github.com/Arthur-Ficial/apfel (fetched 2026-08-28T04:09:41.745356+00:00, sha ff9944028bda)
  - homepage: https://apfel.franzai.com (fetched 2026-08-29T08:42:30.811744+00:00, sha c972dfc8b943)
- Data as of 2026-08-30T08:39:29.467469+00:00.