SeedLinked

On a schedule, the server scrapes the sources and skips records it already has. New records go into a Cloud Tasks queue. The AI step writes structured descriptions covering strengths, weaknesses, growth and yield, flavor and appearance, disease resistance, maturity and growing conditions, and a person approves or rejects them. Approved data is saved to MongoDB and rejected data is removed. If a step fails, the item goes back to the queue and is retried. LangSmith watches each run and can alert Slack through the backend.

When the model, prompt or generation logic changes, a GitHub Actions run executes the offline evals in LangSmith against the AI enrich step. If quality holds, the change ships.

If output quality drops after a model, prompt or logic change, the gate blocks it. The change gets fixed and rerun before anything reaches the live pipeline.

The server scrapes a fixed set of sources and checks source ID and name against the database. Records it already has are ignored, and only new ones move on.

New records are queued through the Cloud Tasks API. If processing fails, the task is retried instead of lost.

A LangChain workflow using Gemini turns the collected data into structured descriptions, coordinated by GCP Workflows.

Generated content is never published directly. A person approves or rejects it first.

Approved seed information is saved to the database and becomes available in the app.

Rejected information is removed from the processing flow and from its associated stored data.

LangSmith covers both kinds of evaluation. Offline, it checks changes before release in CI/CD. Online, it records each production run, so you can see the incoming data and what happened inside the AI enrich step.

LangSmith sends an alert to a webhook on the SeedLinked backend, which turns it into a message for Slack.

The team sees the alert in Slack.

Switching to a new model version, editing a prompt, or changing the generation logic all count as an AI change.

Eval tests run as a GitHub Actions workflow instead of inside the live pipeline, so checking quality never slows it down. The results are recorded in LangSmith.

The regression check: if the generated descriptions get worse, the change is blocked.

Only a change that passes the gate is deployed.