LinkedIn Endpoints no longer work

[flaming910]

Started getting Error 410 on my requests so did some digging and it seems that the endpoints are now in a format such as "voyager/api/identity/dash/profiles/", and the responses might be a bit different too i'm not certain.

[brestok-1]

Did you find a way on how ti fix get_profile method?

[eracle]

It might not be backwards compatible:

pip install jsonpath-ng

    def get_profile(
            self, public_id: Optional[str] = None, urn_id: Optional[str] = None
    ) -> Dict:
        """Fetch data for a given LinkedIn profile.

        :param public_id: LinkedIn public ID for a profile
        :type public_id: str, optional
        :param urn_id: LinkedIn URN ID for a profile
        :type urn_id: str, optional

        :return: Profile data
        :rtype: dict
        """

        member_identity = public_id if public_id else urn_id
        params = {
            'decorationId': 'com.linkedin.voyager.dash.deco.identity.profile.FullProfileWithEntities-91',
            'memberIdentity': member_identity,
            'q': 'memberIdentity',
        }
        res = self._fetch("/identity/dash/profiles", params=params)

        data = res.json()

        if data and "status" in data and data["status"] != 200:
            self.logger.info(f"request failed: {res.content}")
            return {}

        # Now, extract using JSONPath

        # 1. Full Name
        first_name_expr = parse('$.elements[0].multiLocaleFirstName.en_US')
        first_name = first_name_expr.find(data)[0].value if first_name_expr.find(data) else None

        last_name_expr = parse('$.elements[0].multiLocaleLastName.en_US')
        last_name = last_name_expr.find(data)[0].value if last_name_expr.find(data) else None

        full_name = f"{first_name} {last_name}"

        # 2. Headline
        headline_expr = parse('$.elements[0].multiLocaleHeadline.en_US')
        headline = headline_expr.find(data)[0].value if headline_expr.find(data) else None

        # 3. Summary
        summary_expr = parse('$.elements[0].multiLocaleSummary.en_US')
        summary = summary_expr.find(data)[0].value if summary_expr.find(data) else None

        # 4. Location/Address
        address_expr = parse('$.elements[0].multiLocaleAddress.en_US')
        address = address_expr.find(data)[0].value if address_expr.find(data) else None

        country_expr = parse('$.elements[0].geoLocation.geo.defaultLocalizedName')
        country = country_expr.find(data)[0].value if country_expr.find(data) else None

        # 5. Education (list of dicts with key details)
        education_expr = parse('$.elements[0].profileEducations.elements[*]')
        educations_raw = [match.value for match in education_expr.find(data)]
        educations = []
        for edu in educations_raw:
            school = parse('multiLocaleSchoolName.en_US').find(edu)[0].value if parse(
                'multiLocaleSchoolName.en_US').find(edu) else None
            degree = parse('multiLocaleDegreeName.en_US').find(edu)[0].value if parse(
                'multiLocaleDegreeName.en_US').find(edu) else None
            field = parse('multiLocaleFieldOfStudy.en_US').find(edu)[0].value if parse(
                'multiLocaleFieldOfStudy.en_US').find(edu) else None
            start_year = parse('dateRange.start.year').find(edu)[0].value if parse('dateRange.start.year').find(
                edu) else None
            end_year = parse('dateRange.end.year').find(edu)[0].value if parse('dateRange.end.year').find(edu) else None
            educations.append({
                'school': school,
                'degree': degree,
                'field': field,
                'start_year': start_year,
                'end_year': end_year
            })

        # 6. Experience/Positions (list of dicts with key details)
        positions_expr = parse('$.elements[0].profilePositionGroups.elements[*]')
        positions_raw = [match.value for match in positions_expr.find(data)]
        positions = []
        for pos_group in positions_raw:
            title_expr = parse('profilePositionInPositionGroup.elements[0].multiLocaleTitle.en_US')
            title = title_expr.find(pos_group)[0].value if title_expr.find(pos_group) else None

            company_expr = parse('multiLocaleCompanyName.en_US')
            company = company_expr.find(pos_group)[0].value if company_expr.find(pos_group) else (
                parse('company.name').find(pos_group)[0].value if parse('company.name').find(pos_group) else None)

            start_month = parse('dateRange.start.month').find(pos_group)[0].value if parse(
                'dateRange.start.month').find(pos_group) else None
            start_year = parse('dateRange.start.year').find(pos_group)[0].value if parse('dateRange.start.year').find(
                pos_group) else None
            end_month = parse('dateRange.end.month').find(pos_group)[0].value if parse('dateRange.end.month').find(
                pos_group) else None
            end_year = parse('dateRange.end.year').find(pos_group)[0].value if parse('dateRange.end.year').find(
                pos_group) else None

            description_expr = parse('profilePositionInPositionGroup.elements[0].multiLocaleDescription.en_US')
            description = description_expr.find(pos_group)[0].value if description_expr.find(pos_group) else None

            positions.append({
                'title': title,
                'company': company,
                'start': f"{start_month}/{start_year}" if start_month else start_year,
                'end': f"{end_month}/{end_year}" if end_month else end_year,
                'description': description
            })

        # 7. Skills (list of strings)
        skills_expr = parse('$.elements[0].profileSkills.elements[*].multiLocaleName.en_US')
        skills = [match.value for match in skills_expr.find(data)]

        # 8. Certifications (list of dicts)
        certifications_expr = parse('$.elements[0].profileCertifications.elements[*]')
        certs_raw = [match.value for match in certifications_expr.find(data)]
        certifications = []
        for cert in certs_raw:
            name = parse('multiLocaleName.en_US').find(cert)[0].value if parse('multiLocaleName.en_US').find(
                cert) else None
            authority = parse('authority').find(cert)[0].value if parse('authority').find(cert) else None
            certifications.append({'name': name, 'authority': authority})

        # 9. Languages (list of dicts)
        languages_expr = parse('$.elements[0].profileLanguages.elements[*]')
        langs_raw = [match.value for match in languages_expr.find(data)]
        languages = []
        for lang in langs_raw:
            name = parse('multiLocaleName.en_US').find(lang)[0].value if parse('multiLocaleName.en_US').find(
                lang) else None
            proficiency = parse('proficiency').find(lang)[0].value if parse('proficiency').find(lang) else None
            languages.append({'name': name, 'proficiency': proficiency})

        # 10. Volunteer Experiences (list of dicts)
        volunteer_expr = parse('$.elements[0].profileVolunteerExperiences.elements[*]')
        vol_raw = [match.value for match in volunteer_expr.find(data)]
        volunteers = []
        for vol in vol_raw:
            role = parse('multiLocaleRole.en_US').find(vol)[0].value if parse('multiLocaleRole.en_US').find(
                vol) else None
            company = parse('multiLocaleCompanyName.en_US').find(vol)[0].value if parse(
                'multiLocaleCompanyName.en_US').find(vol) else None
            volunteers.append({'role': role, 'company': company})

        # 11. Profile Picture URL (example 200x200)
        pic_root_url_expr = parse('$.elements[0].profilePicture.displayImageReference.vectorImage.rootUrl')
        pic_root_url = pic_root_url_expr.find(data)[0].value if pic_root_url_expr.find(data) else None

        pic_path_expr = parse(
            '$.elements[0].profilePicture.displayImageReference.vectorImage.artifacts[0].fileIdentifyingUrlPathSegment')
        pic_path = pic_path_expr.find(data)[0].value if pic_path_expr.find(data) else None

        profile_pic_url = pic_root_url + pic_path if pic_root_url and pic_path else None

        # 12. Volunteer Causes (list)
        causes_expr = parse('$.elements[0].volunteerCauses[*]')
        causes = [match.value for match in causes_expr.find(data)]

        # Output all extracted info
        extracted_info = {
            'full_name': full_name,
            'headline': headline,
            'summary': summary,
            'address': address,
            'country': country,
            'educations': educations,
            'positions': positions,
            'skills': skills,
            'certifications': certifications,
            'languages': languages,
            'volunteer_experiences': volunteers,
            'profile_pic_url': profile_pic_url,
            'volunteer_causes': causes
        }

        return extracted_info

[davidbernat]

A brief inspection of the View Source and Network appears to show Voyager shifted to a graphql? implementation.

Users at open-linkedin-api began having the same problem.
EseToni/open-linkedin-api#6

If you follow their procedure for finding new API endpoints you see consistently the change in URL form.
https://github.com/nsandman/linkedin-api?tab=readme-ov-file#to-find-endpoints

This is very, very bad for users, and I question their motives at LinkedIn. A thorough search for downloading LinkedIn data repeatedly finds the same three or four useless Reddit.com posts, and not even one browser extension seems to pop up as an alternative for what is obviously a high return use case. The official LinkedIn documents promise "Your LinkedIn data belongs to you, and you can download an archive any time or view the rich media you have uploaded." but on all four occasions that I tried their process over the last two years the results were the same: a few hundred KB of text data which contains none of the images or videos associated with the posts. Their promise to "view the rich media" also leads to a page which contains only the post text with a grey'd out download tab for one image apiece; no organization, no capacity to reconstruct the user posts or timelines. This is very, very bad for users, and mark my words this is the beginning of a new kind of assault of data on the web.

At this point in time zero solutions appear to exist for the most basic cloud-based social network, and I charge intentionally.

I would love to be wrong and hotheaded here but we rowed this boat for two and a half years already. The lists exist. Trust me.

[davidbernat]

While slightly more colorful than my take, users should be reminded that LinkedIn does not allow users to alter the visibility of posts after their posts: in other words, of the thousands of hours of posts that I have created under my copyrighted profile, I am entirely unable to mass-private those posts so that they are not visible. This would be akin to a nuclear option for the solution of changing jobs or privatizing data; and even this remedial step is not afforded by Reid Hoffman and the LinkedIn CEO.

Here is the top result on a popular search engine for the query "linkedin make all posts private". It is NSFW. Ironically.
https://www.reddit.com/r/linkedin/comments/1jnlsr1/you_cant_change_the_visibility_of_a_post_after/

[brestok-1]

@eracle Thank you a lot! I have a strange issue: After each request using your way, I need to login again.. Previous li_at cookie become invalid. Do you know how to solve it?

[eracle]

@davidbernat
Are you a bot? Your post doesn't make much sense to me

@brestok-1
One week after my solution is already wrong. Additionally, this project has outdated login functionality. The login method on this (linkedin-api) library during login uses mobile-based cookies; the get_profile instead does not, so the anti-bot mechanism identifies those types of incongruences. atm, I am working on v2 of another open-source project of mine, but I am forced to use other technologies to perform login (such as Playwright) and then get the cookies from there.
This repo is outdated. I am sorry

[davidbernat]

@eracle The login function appeared to work correctly for me. I do not think the problem is 'anti-bot'. I think the problem is a fundamental restructuring of the Voyager API away from the prior GET URLs, at least in my quick analysis. I did notice that the software often threw cookie outdated errors which were opaque, because browser cookies were up to date, but the package quietly caches cookies in a hidden .linkedin_api directory (usually in ~/) and once that directory is deleted the software will bring in the new correct cookie from the browser. I cannot speak to the other statements you make about the monkeypatches.

@eracle In the jovial sayings of my internal team when we find rude Internet weirdness: "lol. lmao." What about my comment makes no sense? I will happily expound. In bot terms, your posts are more likely to be bots. What would you like me to describe.

[eracle]

@davidbernat No offence was intended. Your comment looked like a bot to Me, and the login function didn't work to Me and apparently also to @brestok-1

[davidbernat]

@eracle No offense was taken. The comment is intended to demonstrate a need to rally between the competent engineers, who are not aware of precisely the dire situation they, and we, are in. And clearly my comment failed. Have a nice day, though.

If anyone solves this problem please create a comment so we can remove out data without third-party driven delays.
Owning our timing is the final frontier, and we possess our own tools. I hope to God we wake up with a different country.

[Mangor1no]

They changed their model from using GraphQL to SSE I think. Been scratching my head off since the update but got no luck so far

[eracle]

https://github.com/eracle/OpenOutreach/blob/master/linkedin/api/client.py
https://github.com/eracle/OpenOutreach/blob/master/linkedin/api/voyager.py

Check this for a working version