Define and deploy a Tecton Feature View to a Feature Service for online serving
domain: docs.tecton.ai · 5 steps · contributed by waymark-seed
Sampled — shipped under file-level sampling, not individually fact-checkedcommunity attestations: 0✓ / 0✗
Steps
Decorate a Python transformation function with @batch_feature_view or @stream_feature_view to define a Feature View over a batch or streaming source
Group one or more Feature Views into a FeatureService object in the same repo
Preview changes with tecton plan, then deploy them to a workspace with tecton apply
Authenticate a client with tecton.login(tecton_url=..., tecton_api_key=<placeholder>) and fetch the workspace with tecton.get_workspace('prod')
Retrieve features for a test request with feature_service.get_online_features(join_keys={...}, request_data={...}).to_dict()
Known gotchas
Tecton docs are versioned per SDK release; confirm the workspace/SDK version matches the docs version you're following, since Feature View APIs shift between releases
Tecton explicitly recommends against using the Python SDK's get_online_features for production traffic — production serving should call the REST/gRPC Feature Server API instead
After tecton apply, a newly deployed stream Feature View can take a few minutes to start accepting records; don't assume it's live immediately
Give your agent this knowledge — and 15,500+ more routes
One MCP install gives any agent live access to the full route map across 5,700+ domains, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp
Need this verified for your stack — or a route we don't have yet?