GoatCounter API

GoatCounter has a rudimentary API; this is far from feature-complete, but solves some common use cases.

The API is currently unversioned and prefixed with /api/v0; breaking changes will be avoided and are not expected but may occur. I'll be sure to send ample notification of this to everyone who has generated an API key.

Authentication

To use the API create a key in your account (Settings → Password, MFA, API); send the API key in the Authorization header as Authorization: bearer [token].

You will need to use Content-Type: application/json; all requests return JSON unless noted otherwise.

Example:

curl -X POST https://example.goatcounter.com/api/v0/export \
    -H 'Content-Type: application/json' \
    -H 'Authorization: Bearer 2q2snk7clgqs63tr4xc5bwseajlw88qzilr8fq157jz3qxwwmz5'

Replace the key and URL with your actual values.

Rate limit

The rate limit is 60 requests per 120 seconds. The current rate limits are indicated in the the headers:

X-Rate-Limit-Limit        Number of requests the rate limit kicks in; this is always the same.
X-Rate-Limit-Remaining    Requests remaining this period.
X-Rate-Limit-Reset        Seconds until the rate limits resets.

Errors

Errors are reported in either an error or errors field; the error field always contains a string; for example:

{
    "error": "oh noes!"
}

The errors field contains an object with a list:

{
    "errors": {
        "key":     ["error1", "error2"],
        "another": ["oh noes!"]
    }
}

A status code in the 2xx range will never contain errors, a status code in the 4xx or 5xx range will always have either error or errors, but never both. There may also be additional data in other fields on errors.

API reference

API reference docs are available at:

Examples

Backend integration

You can use /api/v0/count to send requests from your backend. This is the same as /count but has higher rate-limits, allows setting some additional fields, and allows batching multiple pageviews in one request.

Detail are available in the API reference, a simple example might look like:

#!/bin/sh
token=[your api token] api=https://[my code].goatcounter.com/api/v0 curl() { \command curl \ -H 'Content-Type: application/json' \ -H "Authorization: Bearer $token" \ "$@" } curl -X POST "$api/count" \ --data '{"no_sessions": true, "hits": [{"path": "/one"}, {"path": "/two"}]}'

Export

Example to export via the API:

#!/bin/sh
token=[your api token] api=https://[my code].goatcounter.com/api/v0 curl() { \command curl \ -H 'Content-Type: application/json' \ -H "Authorization: Bearer $token" \ "$@" } # Start a new export, get ID from response object. id=$(curl -X POST "$api/export" | jq .id) # The export is started in the background, so we'll need to wait until it's finished. while :; do sleep 1 finished=$(curl "$api/export/$id" | jq .finished_at) if [ "$finished" != "null" ]; then # Download the export. curl "$api/export/$id/download" | gzip -d break fi done

The above doesn't does no error checking for brevity: errors are reported in the error or errors field as described above.

The export object contains a last_hit_id parameter, which can be used as a pagination cursor to only download hits after this export. This is useful to sync your local database every hour or so:

# Get cursor
start=$(curl "$api/export/$id" | jq .last_hit_id)

# Start new export starting from the cursor.
id=$(curl -X POST "$api/export" --data "{\"start_from_hit_id\":$start}" | jq .id)