matrx-ai 0.1.7

Matrx AI Engine — unified AI orchestration backend with multi-provider LLM support, streaming, tool execution, and conversation persistence.

matrx-ai

Matrx AI Engine — unified AI orchestration backend with multi-provider LLM support, streaming, tool execution, and conversation persistence.

Installation

# Core library only (AI providers, tools, conversation management)
pip install matrx-ai

# With the FastAPI server layer
pip install "matrx-ai[server]"

# Or with uv
uv add matrx-ai
uv add "matrx-ai[server]"

Quick Start (library)

import matrx_ai

# Initialize once at startup (registers the database connection)
matrx_ai.initialize()

# Use the AI engine directly
from matrx_ai.orchestrator import execute_ai_request
from matrx_ai.config.unified_config import UnifiedConfig

Quick Start (server)

uv sync --extra server   # install with server deps
cp .env.example .env     # fill in your API keys
make dev                 # start dev server on :8000

Environment Variables

Copy .env.example to .env and fill in the required keys:

• AI provider keys: OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, GROQ_API_KEY, etc.
• Supabase: SUPABASE_URL, SUPABASE_ANON_KEY, SUPABASE_SERVICE_ROLE_KEY, SUPABASE_MATRIX_*

Common Commands

make dev        # dev server with reload on :8000
make run        # production mode (matrx-ai-server)
make lint       # ruff check
make fmt        # ruff format
make typecheck  # pyright
make test       # pytest -v

Publishing a New Release

./scripts/publish.sh              # patch bump  (e.g. 1.2.3 → 1.2.4)
./scripts/publish.sh --minor      # minor bump  (e.g. 1.2.3 → 1.3.0)
./scripts/publish.sh --major      # major bump  (e.g. 1.2.3 → 2.0.0)
./scripts/publish.sh --message "feat: add new provider"
./scripts/publish.sh --dry-run    # preview without committing

Version History

Version	Highlights
v0.1.7	Patch release
v0.1.6	Patch release
v0.1.5	Patch release
v0.1.4	Patch release
v0.1.3	Patch release
v0.1.2	Patch release
v0.1.1	Patch release
v0.1.0	Initial release — multi-provider AI orchestration, streaming, tool system, conversation persistence