Submit search result feedback

Tell Pearch whether a profile returned by a search was right, and why. Pearch uses
feedback to measure and improve search quality.

Identify the result with the /v2/search response's thread_id and the profile's
search_results[].profile.linkedin_slug. You can send feedback at any time later:
search_id (the response's uuid) is optional and defaults to the thread's latest
completed search.

Three levels; send as much as you have:

  1. Rating. rating is up, down or maybe. This is all a thumbs button needs.
  2. Profile feedback: why. profile_feedback is a string or a list. A string is free
    text, for example "15 years of experience, we need 3-5". An object carries a
    code: a preset from GET /v2/search/feedback/reasons when one fits, otherwise your
    own code, up to three codes per profile. Add unstated: true when the criterion was
    not in the search request, such as an employer you exclude.
  3. Requirement feedback: which requirements. requirement_feedback gives a verdict
    for each requirement you checked, by its requirement_index. A recruiter who marks
    the requirements a profile misses sends does_not_meet for each of them.

Only requirement feedback refers to requirements. requirement_index is the
position of the requirement in the /v2/search response's search_requirements,
where every entry carries its requirement_index. Searches made from a
natural-language query list the requirements Pearch parsed from it. You never send
requirement text.

Each submission replaces the stored feedback for that (thread_id, linkedin_slug),
whoever sends it, so send the rating, profile feedback and requirement feedback
together. To send many decisions at once, for example from an ATS, use
POST /v2/search/feedback/batch.

A rejected request returns 400 or 404 with the reason in status, for example
'rating': rating must be 'up', 'down' or 'maybe'. String values are trimmed; rating,
author_type and codes are case-insensitive. Unknown fields are rejected. Older names
are still accepted: chat_thread_id and search_call_uuid for thread_id and
search_id, reasons and verdicts for profile_feedback and requirement_feedback,
and a top-level comment or reason as a free-text profile feedback item.

Body Params
uuid
required

The thread_id returned by /v2/search. chat_thread_id is accepted as its older name.

string
required

The rated profile's search_results[].profile.linkedin_slug value.

uuid | null

The uuid returned by /v2/search. Optional; defaults to the thread's latest completed search. search_call_uuid is accepted as its older name.

string
enum
required

maybe marks a borderline profile.

Allowed:
length ≤ 500

Why the profile fits or not, as one string or a list; up to three items with a code. reasons is accepted as its older name.

requirement_feedback
array of objects
length ≤ 50

Verdicts on the requirements you checked; each requirement at most once. verdicts is accepted as its older name.

requirement_feedback
string
enum
Defaults to person

Who produced the rating; mixed for a person confirming an agent's rating.

Allowed:
Responses

400

Invalid request: an invalid value, a preset code that does not match the rating, or a requirement index outside the search's requirements. status names the field and the rule.

401

Invalid API key

403

Invalid API key

404

No completed search in the thread (or none with this search_id), or the profile is not in that search's ranking. status says which.

Language
Credentials
Bearer
JWT
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json