philosopher-ai
v1.0.1
Published
Socratic AI Terminal Coding Assistant & Multi-Agent Orchestrator
Maintainers
Readme
🏛️ Philosopher AI (Filozof)
Islamic Golden Age Wisdom × Modern Software Engineering × Autonomous AI
Philosopher AI is a state-of-the-art terminal-native AI coding assistant and multi-agent orchestrator. Featuring a distinct persona inspired by the scholars of the Islamic Golden Age, a modern Terminal UI rendered in vibrant light orange (#FF7A00) and crisp white, smart repo mapping, customizable skills, and specialized scholar sub-agents.

🌟 Key Features
🏛️ Meclis & Scholarly Multi-Agent Council:
- Meclis (Orchestrator): Council Leader. Decomposes tasks, delegates work, and synthesizes clean code.
- Farabi (Architect): "Önce problemi ve sistemi anla, sonra kodu yaz." System design, clean architecture, and planning.
- Biruni (Researcher): "Bir iddiayı kabul etmeden önce ölç." Empirically investigates the codebase using read-only search and AST file tree mapping.
- Sina (Diagnostics): "Belirtiyi değil, hastalığın kaynağını tedavi et." Bug finding, root-cause analysis, and safe fixes.
- Khwarizmi (Algorithm): "Problemi doğru modele dönüştür." Algorithms, optimization, and modeling.
- Haldun (Systems): "Bir sistem yalnızca bugün değil, zaman içinde nasıl davranacak?" Large codebase evolution, technical debt tracking, and scaling.
🎨 Modern Design & Styling:
- Light Orange (
#FF7A00) & White brand identity with elegant boxes and spinners. - Interactive Unified Diff preview for all code edits.
- Styled terminal Markdown renderer.
- Light Orange (
🛠️ Multi-Model & Provider Support:
- Anthropic: Claude 3.5 Sonnet / Claude 3 Opus
- OpenAI: GPT-4o / GPT-4o-mini
- Google Gemini: Gemini 2.0 Flash / Pro
- Ollama: Local open-weights models (Qwen2.5-Coder, DeepSeek-R1)
🔒 Safety & Permission Guard:
- Interactive Mode: Prompts user confirmation before running shell commands or mutating files.
- YOLO Mode: Seamless auto-execution mode.
📚 Extensible Skills Engine:
- Custom skills loaded automatically from
.philosopher/skills/or.filozof/skills/.
- Custom skills loaded automatically from
🚀 Quick Start & Installation
Requirements
- Node.js ≥ 22 (native SQLite module requirement)
Global Install (recommended)
npm install -g philosopher-ai
# Launch Philosopher Interactive Terminal REPL in any project
cd your-project
philosopher
# Or run non-interactively
philosopher -p "Refactor the database connection to use async pooling"From Source
git clone <repository-url>
cd Philosopher
# Install dependencies
npm install
# Build TypeScript
npm run build
# Launch Philosopher Interactive Terminal REPL
npm start
# Or run non-interactively
npx tsx src/index.ts -p "Refactor the database connection to use async pooling"Configuration
Copy .env.example to .env and add at least one provider API key:
cp .env.example .envCustomizing Rules & Personas per Project
Philosopher ships with default council rules (safety, reasoning, coding, council) and scholar personas as bundled templates. Any project can override them by creating:
your-project/
├── .philosopher/
│ ├── rules/ # overrides ALL bundled council rules
│ └── context.md # task execution context
└── agents/
└── farabi/system.md # overrides one persona (etc.)💬 Slash Commands
| Command | Description |
| :--- | :--- |
| /help | Display guidance manual |
| /agents | Inspect active scholarly council (Meclis, Farabi, Biruni, Sina, Khwarizmi, Haldun) |
| /skills | View loaded custom skills (.philosopher/skills/) |
| /plan | Formulate an implementation plan (Farabi) |
| /parallel | Parallel DAG execution via PlannerEngine |
| /map | Generate codebase file tree and symbol map |
| /mode | Toggle between Interactive (interactive) and Auto (yolo) mode |
| /model | Switch active provider/model |
| /config | View active LLM providers and API keys |
| /memory | Show hierarchical memory status |
| /session | Session management (list/load/new) |
| /resume | Resume a past conversation |
| /sessions | Manage saved sessions |
| /new | Start a new session |
| /copy | Copy last agent response to clipboard |
| /retry | Retry last exchange |
| /undo | Undo last file changes (checkpoint restore) |
| /export | Export chat as Markdown |
| /clear | Reset conversation history |
| /tdd | Test-Driven Development cycle (Sina+Khwarizmi) |
| /refactor | Project-wide refactor (Haldun) |
| /review | Review git diff (Sina+Khwarizmi) |
| /commit | Smart commit message generation |
| /blame | Git blame analysis |
| /research | Deep research (Biruni) |
| /debate | Scholar debate on a topic |
| /gnostic | Gnostic/chaos mode toggle |
| /add | Add file to context |
| /paste | Paste from clipboard |
| /branch | Create git branch |
| /bump | Version bump |
| /diagnose | Diagnose an error (Sina) |
| /train | Train custom skill from history |
| /help | Show all commands |
| /exit | Exit Philosopher session |
| Keys | Tab cycle views · Ctrl+P palette · Ctrl+F search · Ctrl+O code expand · Ctrl+G bottom · PgUp/Dn scroll · ? help |
📄 License
MIT — see LICENSE
