An ActivityPub-compatible server and library built with FastAPI. Provides storage, processing, and HTTP endpoints for Activity Streams 2.0 objects — all in a single importable package.

activity_serve/
├── core/       # JSON-LD processing and utilities
├── store/      # Object storage with pluggable backends
├── bus/        # Rule-based activity processing engine
├── api/        # FastAPI HTTP endpoints
├── services/   # User management, authentication
└── middleware/  # Request logging, normalization

Use as a library:

from activity_serve.store import ActivityStore
from activity_serve.bus import ActivityBus, when
from activity_serve.core.ld import frame
from activity_serve.core.utils import first_id

Use as an app:

from activity_serve.main import create_app

app = create_app()

uv add activity-serve

# With optional backends
uv add "activity-serve[es]"     # Elasticsearch storage
uv add "activity-serve[redis]"  # Redis caching

For development:

git clone https://github.com/yourusername/activity-serve.git
cd activity-serve
uv sync

Environment variables:

GOOGLE_CLIENT_ID=your-google-client-id

For Elasticsearch storage:

ELASTICSEARCH_URL=http://localhost:9200

For Redis caching:

REDIS_URL=redis://localhost:6379/0

uv run uvicorn activity_serve.main:app --reload

uv run pytest

# Integration tests (require running services)
ELASTICSEARCH_URL=http://localhost:9200 uv run pytest tests/store/test_elasticsearch_backend.py
REDIS_URL=redis://localhost:6379/0 uv run pytest tests/store/test_redis_cache.py

• GET /me — Current authenticated user
• GET /u/{key}/inbox — User inbox
• GET /u/{key}/outbox — User outbox
• POST /u/{key}/outbox — Submit an activity
• GET /u/{key}/{path} — Query any sub-path (objects, collections)
• GET /admin — Admin UI
• GET /healthz — Health check

ActivityStore manages storage and caching of Activity Streams objects. Backends are pluggable — in-memory for development, Elasticsearch for production, Redis for caching.

ActivityBus processes activities through a rule-based behavior system. Register behaviors with the @when decorator to match activities by JSON-LD pattern:

from activity_serve.bus import when

@when({"type": "Create", "object": {"type": "Note"}})
def handle_new_note(activity):
    print(f"New note created: {activity['object']['content']}")

core.ld provides JSON-LD operations — normalize, expand, compact, and frame — built on pyld.

MIT