Docs / Quickstart

Quickstart

Send your first event and see your first root cause in about five minutes.

1. Create a project and get an API key

  1. Sign up — no card required.
  2. A project and an ingest key are created for you automatically.
  3. Copy your key from Settings → API Keys. It's shown once.
export SCRYWATCH_API_KEY="sk_live_xxxxxxxxxxxxxxxx"

2. Send your first event

One HTTPS call. No agent, no SDK required.

curl -X POST https://api.scrywatch.com/api/ingest \
  -H "Authorization: Bearer $SCRYWATCH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "events": [
      {
        "level": "error",
        "type": "log",
        "message": "Checkout failed: payment provider timeout",
        "service": "checkout-api",
        "environment": "production",
        "session_id": "sess_8c2f",
        "user_id": "user_4471",
        "metadata": { "order_id": "ord_99213", "provider": "stripe" }
      }
    ]
  }'

You'll get a 202 Accepted. Your event is queued and appears in seconds.

{ "accepted": 1 }

Field reference

Field Required Notes
level yes debug · info · warn · error
type yes log · crash · event
message yes The human-readable line. Fingerprinted automatically.
service – e.g. checkout-api. Powers facets and Deploy Diff.
environment – e.g. production, staging.
session_id – Ties events into a replayable session.
user_id – Filter and group by user.
timestamp – ISO 8601 or epoch ms. Defaults to receive time.
metadata – Any JSON object. Searchable.

Send up to 50 events per request. Just keep posting batches.

3. See it land

Open Search. Filter by service: checkout-api and notice it's already grouped into a pattern — the variable bits (order IDs, timeouts) are tokenized so the same failure always groups together.

4. Find your first root cause

This is the point of ScryWatch.

Mark a deploy — from the dashboard or your CI:

curl -X POST https://api.scrywatch.com/api/orgs/<org>/projects/<project>/deploys \
  -H "Authorization: Bearer $SCRYWATCH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "service": "checkout-api", "environment": "production", "version": "v1.4.2" }'

Send a few more errors, then open Deploy Diff for v1.4.2. You'll see production behavior before vs. after — the new error pattern, how often it's firing, and the issue it rolled up into.

Deploy Diff comparing before and after v1.4.2

5. Connect your AI assistant (optional)

Let Claude or Cursor query everything above, read-only. Connect your assistant →

Next steps