Attribution report (organic/paid, first/last touch)

Also reachable by the internal remote-MCP server as get_attribution_report: the handler uses requireProjectRoleOrMcpScope(VIEWER, ‘mcp:attribution_read’). MCP access is mediated by a separate internal worker (x-worker-token), not by a directly-presented customer API key, so ApiKeyAuth is NOT advertised here.

Authorizations

SessionCookie

Parameters

Path Parameters

projectId

required

string

Query Parameters

startDate

string format: date

Inclusive window start (YYYY-MM-DD). Parsed with parseDateRangeParams; when either startDate or endDate is missing/invalid the route falls back to a rolling default window (30 days for events/visits/attribution, 7 days for growth).

endDate

string format: date

Inclusive window end (YYYY-MM-DD). See startDate for fallback behavior.

model

string

default: first

Attribution model (default “first”).

window

integer

default: 30

Attribution lookback window in days.

Responses

200

Attribution breakdown (dynamic shape). organic/paid touch entries are {source, count}; campaign entries add platform (the utmSource half of the “campaign::source” composite key, null when empty). Each list is capped at 10 entries, sorted by count DESC. _meta.model and _meta.window are pass-through echoes of the model / window query params and are NOT used to change the computation — window is Number(...) so it can be NaN, which JSON.stringify serializes as null. organic.* sources are classifier channel labels (e.g. “organic:google”) or the literal “直接访问”; paid.* sources are human labels from PAID_CLICK_ID_LABELS (Google Ads, Meta, 百度…) falling back to the raw click-ID key.

application/json

object

organic

required

object

firstTouch

required

Array<object>

<= 10 items

object

source

required

string

count

required

integer

lastTouch

required

Array<object>

<= 10 items

object

source

required

string

count

required

integer

paid

required

object

firstTouch

required

Array<object>

<= 10 items

object

source

required

string

count

required

integer

lastTouch

required

Array<object>

<= 10 items

object

source

required

string

count

required

integer

campaign

required

object

firstTouch

required

Array<object>

<= 10 items

object

source

required

Utm_campaign, or the literal “unknown”

string

platform

required

Utm_source of the touch; null when empty

string | null

count

required

integer

lastTouch

required

Array<object>

<= 10 items

object

source

required

Utm_campaign, or the literal “unknown”

string

platform

required

Utm_source of the touch; null when empty

string | null

count

required

integer

_meta

required

object

queryDurationMs

required

integer

model

required

Echo of the model query param; defaults to “first”. Not applied to the computation.

string

window

required

Echo of the window query param coerced with Number(); defaults to 30. Serializes as null if the param is non-numeric (NaN).

number | null

rowsAnalyzed

required

Number of visit AE rows fetched for the range

integer

GET

/api/projects/{projectId}/analytics/attribution

1
curl --request GET \
2
  --url 'https://retidal.com/api/projects/%3Cstring%3E/analytics/attribution?model=first&window=30' \
3
  --cookie trackly_session=%3CYOUR_API_KEY%3E

1
const url = 'https://retidal.com/api/projects/%3Cstring%3E/analytics/attribution?model=first&window=30';
2
const options = {method: 'GET', headers: {cookie: 'trackly_session=%3CYOUR_API_KEY%3E'}};
3

4
try {
5
  const response = await fetch(url, options);
6
  const data = await response.json();
7
  console.log(data);
8
} catch (error) {
9
  console.error(error);
10
}

1
import requests
2

3
url = "https://retidal.com/api/projects/%3Cstring%3E/analytics/attribution"
4

5
querystring = {"model":"first","window":"30"}
6

7
headers = {"cookie": "trackly_session=%3CYOUR_API_KEY%3E"}
8

9
response = requests.get(url, headers=headers, params=querystring)
10

11
print(response.json())

1
import httpx
2

3
url = "https://retidal.com/api/projects/%3Cstring%3E/analytics/attribution"
4

5
params = [("model", "first"), ("window", "30")]
6

7
cookies = {"trackly_session": "<YOUR_API_KEY>"}
8

9
response = httpx.get(url, params=params, cookies=cookies)
10

11
print(response.json())

1
package main
2

3
import (
4
  "fmt"
5
  "net/http"
6
  "io"
7
)
8

9
func main() {
10

11
  url := "https://retidal.com/api/projects/%3Cstring%3E/analytics/attribution?model=first&window=30"
12

13
  req, _ := http.NewRequest("GET", url, nil)
14

15
  req.Header.Add("cookie", "trackly_session=%3CYOUR_API_KEY%3E")
16

17
  res, _ := http.DefaultClient.Do(req)
18

19
  defer res.Body.Close()
20
  body, _ := io.ReadAll(res.Body)
21

22
  fmt.Println(res)
23
  fmt.Println(string(body))
24

25
}

200Attribution breakdown (dynamic shape). organic/paid touch entries are {source, count}; campaign entries add platform (the utmSource half of the “campaign::source” composite key, null when empty). Each list is capped at 10 entries, sorted by count DESC. _meta.model and _meta.window are pass-through echoes of the model / window query params and are NOT used to change the computation — window is Number(...) so it can be NaN, which JSON.stringify serializes as null. organic.* sources are classifier channel labels (e.g. “organic:google”) or the literal “直接访问”; paid.* sources are human labels from PAID_CLICK_ID_LABELS (Google Ads, Meta, 百度…) falling back to the raw click-ID key.

1
{
2
  "organic": {
3
    "firstTouch": [
4
      {
5
        "source": "example",
6
        "count": 1
7
      }
8
    ],
9
    "lastTouch": [
10
      {
11
        "source": "example",
12
        "count": 1
13
      }
14
    ]
15
  },
16
  "paid": {
17
    "firstTouch": [
18
      {
19
        "source": "example",
20
        "count": 1
21
      }
22
    ],
23
    "lastTouch": [
24
      {
25
        "source": "example",
26
        "count": 1
27
      }
28
    ]
29
  },
30
  "campaign": {
31
    "firstTouch": [
32
      {
33
        "source": "example",
34
        "platform": "example",
35
        "count": 1
36
      }
37
    ],
38
    "lastTouch": [
39
      {
40
        "source": "example",
41
        "platform": "example",
42
        "count": 1
43
      }
44
    ]
45
  },
46
  "_meta": {
47
    "queryDurationMs": 1,
48
    "model": "example",
49
    "window": 1,
50
    "rowsAnalyzed": 1
51
  }
52
}