Release notes archive

Version 9.16.7🔗︎

December 4, 2023

Changes🔗︎

Preparing for rebrand to 'Tx Platform'. Users do not need to make any changes. No disruption in service is expected.

Version 9.16.6🔗︎

November 30, 2023

Changes🔗︎

Preparing for rebrand to 'Tx Platform'. Users do not need to make any changes. No disruption in service is expected.

Version 9.16.4🔗︎

November 16, 2023

New Features🔗︎

CV/Resume Parser FlexRequests (Beta)🔗︎

Powered by our LLM Parsing Engine, all CV/Resume Parser requests can now include FlexRequests. FlexRequests allow you to ask any question about a resume. More details can be found here.

Improvements🔗︎

CV/Resume & Job Parser🔗︎

For Skills V1, removed dangerous two-letter skills such as AP (Accounts Payable or Andhra Pradesh) and AR (Accounts Receivable or Arkansas).

Do not output a language competency when the CV contains things like serviced Dutch clients, etc.

No longer output country name as part of the work history description in certain cases.

Recognize National Senior Certificate as an educational degree.

Version 9.16.3🔗︎

October 26, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed bug that would cause a 500 status code to be returned when parsing some Hebrew, Japanese, and Turkish documents.

Version 9.16.2🔗︎

October 25, 2023

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Increased maximum file size limit to 16MB.

LLM Parsing Engine (Beta)🔗︎

Greatly improved the Professional Summary using the power of our LLM Parsing Engine.

Now giving more clear error message returned alongside 429 status codes.

Version 9.16.1🔗︎

October 12, 2023

Improvements🔗︎

LLM Parsing Engine (Beta)🔗︎

Increased amount of skills found during parsing.

Bug Fixes🔗︎

LLM Parsing Engine (Beta)🔗︎

Fixed issue causing duplicate Profession classes to be output.

Fixed issue not correctly calculating total months experience and last used date for skills.

Version 9.16.0🔗︎

October 2, 2023

New Features🔗︎

CV/Resume Parser🔗︎

The LLM Parsing Engine is now available as a beta feature. It can be enabled by setting the UseLLMParser property on resume parsing requests. More details can be found here.

Version 9.15.3🔗︎

September 7, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Added many new V2 Skills.

Version 9.15.2🔗︎

August 25, 2023

Bug Fixes🔗︎

All Components🔗︎

Stability improvements.

Version 9.15.1🔗︎

August 24, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Improved parsing for rare edge cases in LinkedIn profiles, such as when a company name and dates are split across pages.

Version 9.15.0🔗︎

August 15, 2023

New Features🔗︎

Skills Intelligence APIs🔗︎

Suggest Skills From Skills- given a set of skills, suggest more similar skills.

Skills Similarity Score- given two sets of skills, calculate the similarity score between the skill sets.

All Ontology API endpoints now return skill descriptions in any supported language when you set OutputLanguage.

Improvements🔗︎

Search & Match🔗︎

Search now supports results past 1000 documents. For more information, reach out to sales@textkernel.com.

Improved ranking of search results when using skill:() or title:() clauses in the SearchExpression field.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed an issue with Portuguese LinkedIn PDF work history parsing.

Version 9.14.8🔗︎

August 1, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Skills normalization is now supported for Turkish CVs.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

In V9, fixed skills output for Hebrew, Japanese, and Turkish CVs when the skill was not found in a particular section of the CV.

Corrected the Profession Group IDs to match the values used in the Skills Intelligence APIs.

Job Parser🔗︎

Corrected the Profession Group IDs to match the values used in the Skills Intelligence APIs.

Version 9.14.7🔗︎

July 19, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Improved certification parsing when the description includes the word 'Certification'.

Better city/state/country parsing when no postal code or street address is included, especially for locations in US and India.

Now recognize VTU, VIT, and ISB as universities.

Other minor improvements for education and location parsing in India.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed issue in LinkedIn PDFs where some descriptions of time periods could be mistaken for a work history entry.

Version 9.14.6🔗︎

June 30, 2023

Bug Fixes🔗︎

All Components🔗︎

Stability improvements.

Version 9.14.5🔗︎

June 21, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Fixed bug when parsing education on Hebrew, Japanese, and Turkish documents.

Version 9.14.3🔗︎

June 7, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed bug where Employer Locations were not being output on Hebrew, Japanese, and Turkish documents.

Minor fix for LinkedIn profiles with very long company names.

Version 9.14.2🔗︎

May 25, 2023

Bug Fixes🔗︎

Job Parsing🔗︎

Fixed bug where sending images caused an unhandled exception. Now the API response will be 422 - Unprocessable Entity.

Version 9.14.1🔗︎

May 17, 2023

New Features🔗︎

Download Indexed Documents🔗︎

You can now download a list of all documents in your indexes via the Tx Console. Review the Standard Transaction Costs for pricing. The download includes 2 columns - the document id and the document's revision date.

Bug Fixes🔗︎

Job Parsing🔗︎

Fixed bug where the Required field was always output as false when using V2 Skills.

Skills Intelligence APIs🔗︎

We now give a friendlier error message when passing invalid ProfessionCodeIds to the Suggest Skills endpoint.

Version 9.14.0🔗︎

May 11, 2023

Improvements🔗︎

Job Parsing Output🔗︎

Significantly improved job parsing accuracy and greatly reduced error rates for almost all output fields.

Added many new fields to the parsing output. Some languages do not include these fields in the output. See this document for additional details.

Version 9.13.5🔗︎

May 4, 2023

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Added many new V2 Skills.

V2 Skills are now available in the following additional languages: Croatian, Czech, Greek, Hungarian, Romanian, Russian, Slovak, Slovenian, and Turkish.

Skills Intelligence APIs🔗︎

The Skills APIs are now available in the following additional languages: Croatian, Czech, Greek, Hungarian, Romanian, Russian, Slovak, Slovenian, and Turkish.

Version 9.13.4🔗︎

April 14, 2023

Improvements🔗︎

Search & Match🔗︎

Minor stability improvements during indexing.

Version 9.13.3🔗︎

April 13, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Minor stability improvements.

Version 9.13.2🔗︎

March 30, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Major improvements to candidate name parsing, especially in cases where the name appears in an email address or LinkedIn URL.

Polish LinkedIn CVs now parse correctly.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed an issue where sometimes GPA would be picked up as a date in the education section.

Some minor runtime bug fixes and rare timeout fixes.

Search & Match🔗︎

Fixed the 'check if document exists' endpoint to return 404 - Not Found if the document does not exist.

Version 9.13.1🔗︎

March 9, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Decreased time to timeout parsing on Turkish, Hebrew, and Japanese documents. A timeout in these languages will return a 408 - Request Timeout response.

Skills Intelligence APIs🔗︎

Autocomplete and Lookup endpoints now give friendlier error messages when using invalid input and output languages.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Search & Match🔗︎

Fixed bug giving 500 when trying to index a resume without metadata. A 400 status code will now be returned instead.

Version 9.13.0🔗︎

March 2, 2023

New Features🔗︎

Skills Intelligence APIs🔗︎

New APIs focused on skills, professions, and the relationship between skills and professions. View more info here.

Transaction Cost Updates🔗︎

This release introduces an improvement in the transaction cost structure for accounts with Search & Match enabled. Following this change, you will be able to parse more documents using the same number of credits, when you choose not to index all of the parsed documents.

The transaction cost structure will be changed as follows:

Action	Old base transaction cost	New base transaction cost
Parse resume or job	2 credits when matching is enabled, else 1	1 credit
Index or update a full resume or job	0 credits	1 credit
Update tags for an already indexed document	0 credits	0 credits

Please note that the base transaction cost does not cover additional services such as geocoding and normalization of professions and skills. The cost for these add-ons will remain unchanged. You can see full details regarding transaction costs here.

Improvements🔗︎

Search & Match🔗︎

Documents parsed without Search & Match enabled can now later be indexed and used for matching/searching.

CV/Resume Parser🔗︎

Improved name parsing on English resumes from Indian locales.

Bug Fixes🔗︎

Search & Match🔗︎

Fixed bug sometimes outputting NaN for SuggestedCategoryWeights on bimetric scoring responses.

Version 9.12.3🔗︎

February 16, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

For Portuguese resumes, correctly output Portuguese as the 'mother tongue' instead of Spanish.

Hebrew, Japanese and Turkish resumes now correctly output the FoundMonth and FoundDay fields on Dates.

Fixed various issues where parsing didn't properly timeout.

Other minor runtime bug fixes.

Version 9.12.1🔗︎

February 1, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Stability improvements.

Version 9.12.0🔗︎

January 26, 2023

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

New Features🔗︎

Skills Intelligence APIs - Beta🔗︎

This release included a beta launch of the Skills Intelligence APIs. The full version was released as part of the 9.13.0 release. View more info here.

Version 9.11.4🔗︎

January 18, 2023

Improvements🔗︎

CV/Resume Parser🔗︎

Improved accuracy for parsing names/locations when the language is English and the locale is India.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Output correct type for certain web addresses, such as GitHub.

For Swedish resumes, do not output Danish as the 'mother tongue'.

Fixed some rare runtime bugs in the parser.

Job Parser🔗︎

Do not consider languages found in certain sections of a job order as 'required'.

Search & Match🔗︎

Fixed an issue where specifying geocoordinates during job parsing caused an error.

Handle certain whitespace in index names instead of causing an error.

Resumes with a very large number of skills no longer cause an error during matching.

Version 9.11.2🔗︎

December 15, 2022

Bug Fixes🔗︎

Document Converter🔗︎

Fixed an issue where text files would not generate HTML when OutputHtml was set to true.

Version 9.11.1🔗︎

December 14, 2022

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed regression with LinkedIn Profile PDF parsing that was introduced in v9.11.0.

Version 9.11.0🔗︎

December 12, 2022

New Features🔗︎

CV/Resume Parser🔗︎

Japanese and Turkish are now supported languages. See this document for additional details.

CV/Resume & Job Parser🔗︎

Significantly improved multi-column and PDF text conversion resulting in general parsing improvements across the board.

Added version selection to professions normalization for taxonomies that offer multiple versions. ONET "2010" is deprecated. Use "2019" instead. See ONET Version for more details.

Improvements🔗︎

CV/Resume Parser🔗︎

Major improvements for Hebrew parsing.Low usage fields are no longer output for this language.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Fixed various issues where parsing didn't properly timeout.

Version 9.10.4🔗︎

November 28, 2022

Improvements🔗︎

CV/Resume Parser🔗︎

Improved recognition of university names.

Report Email-myname@company.com as myname@company.com instead of Email-myname@company.com and related scenarios.

Version 9.10.3🔗︎

November 22, 2022

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed runtime parsing error.

Fixed various scenarios that rarely caused requests to take more than 25s.

Version 9.10.1🔗︎

November 10, 2022

Bug Fixes🔗︎

Search & Match🔗︎

Fixed error where unrelated job titles were being considered a variation of each other. This was only an issue in the matching logic and had no impact on parsing and doesn't require any reindexing.

Version 9.10.0🔗︎

November 03, 2022

Improvements🔗︎

CV/Resume Parser🔗︎

Improve name detection when split across multiple lines

Added support for NIH Stroke Scale certification

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed detection of employment history when denoted by Current Employer header.

Reduce reporting of false positive company names when a company isn't specified.

Reduce false positive certifications from being reported.

Minor runtime bug fixes.

Version 9.9.1🔗︎

October 19, 2022

Bug Fixes🔗︎

Job Parser🔗︎

Fixed bug ouputting V2 Skills in Version 10.

SOAP (deprecated)🔗︎

Fixed an issue that would return an invalid URL when generating a new service proxy.

Fixed bug that would return errors when using SOAP version 1.2.

Version 9.9.0🔗︎

October 11, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Speed improvements.

Added many new V2 Skills.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Version 9.8.1🔗︎

August 3, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better LinkedIn URL parsing when URL is split across multiple lines.

Improved parsing of Danish LinkedIn PDFs.

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Fixed redaction of candidate first name from candidate summary when using V2 skills.

Search & Match🔗︎

Fixed error that would occur when filtering by Job Titles and including a null Settings parameter in the request.

Version 9.8.0🔗︎

July 18, 2022

Enhancements🔗︎

CV/Resume & Job Parser🔗︎

Major upgrade of the Skills Taxonomy (V2)🔗︎

Upgrade for freeto Skills V2. Benefits of V2 include:
- 2x larger skills taxonomy, updated frequently based on real-world data.
- 15-40% higher accuracy of extracted skills.
- Ability to use Skills Normalization (see below).
- When using V2, the candidate summary will no longer contain skills information unless Skills Normalization is also enabled (see below)

Enable Skills Normalization to benefit from better skills classification🔗︎

New! Skill types are output (IT/Professional/Soft Skills)
Enhances the candidate summary with structured skills information
Better clustering of skill synonyms
Compatibility with the taxonomy used in Textkernel's Skills Intelligence APIs and Jobfeed, enabling standardization of taxonomies across all of your data and benchmarking against jobs posted online.
Consumes 0.1 additional credits per transaction

New Features🔗︎

CV/Resume & Job Parser🔗︎

Normalization of Job Titles🔗︎

Enable Professions Normalization to standardize job titles according to a taxonomy.

For Resumes, the most recent 3 job titles are normalized. View the new output properties here.
Supported Taxonomies: Textkernel, ISCO 2008, O*NET 2010. ONET 2019 will be added soon.
The professions taxonomy and the mappings are compatible with the taxonomies used in Textkernel's Skills Intelligence APIs and Jobfeed, enabling standardization of taxonomies across all of your data and benchmarking against jobs posted online.
Supported languages: English, German, French, Spanish, Italian, Dutch, Japanese, Simplified Chinese, Portuguese, Polish
Consumes 0.2 additional credits per transaction

Search & Match🔗︎

Normalization of Job Titles🔗︎

Get more search and match results by extending the query beyond exact job titles.

Normalized job titles are automatically included in Match queries
- Requires the candidates and jobs to be parsed and indexed with Professions Normalization enabled
- Manual search queries can be normalized in real-time through the API (see here)
- Consumes 0.2 additional credits per transaction

Version 9.7.5🔗︎

May 4, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Added some new IT skills.

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Fixed a few minor runtime errors.

Search & Match🔗︎

Better handling of invalid Index and Document IDs.

Version 9.7.4🔗︎

April 6, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Added many new IT skills.

Removed some job titles from the skills data. Users should use the Configuration parameter to output job titles as skills.

Small improvement to Awards section header recognition.

Bug Fixes🔗︎

Search & Match🔗︎

Fixed SearchExpression validation that caused occasional search expressions to return a bad request. This occured when the search expression was only a preposition and used the title:() syntax.

Version 9.7.3🔗︎

March 23, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better detection/cleanup of HTML-encoded entities in text that is only partially HTML.

Small improvement to Certifications section header recognition.

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Fixed a problem where some LinkedIn profiles would be truncated on certain words/phrases.

Matching UI🔗︎

Do not require the user to login if the email address (Username) contained uppercase characters.

Version 9.7.2🔗︎

February 10, 2022

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Improved municipality and region parsing in all languages/locales.

Better date alignment, and therefore work history parsing, for some LinkedIn profiles with 2+ positions at one employer.

Improved de-columnization for LinkedIn profiles with only 1 position listed. This greatly improves parsing for these profiles.

Resume Formatter🔗︎

Skill tags are now a neutral color.

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Don't allow Japanese punctuation in an English resume to cause the language to be reported as Japanese.

[v10 only] Fixed a mapping issue that was causing a DateOfBirth of 0001-01-01 to be output when no birth date was found.

Eliminated rare cases where setting OutputHtml = true was causing the parsing request to hang.

Resume Formatter🔗︎

No longer output company email/phone/etc if those are not provided in the API request.

Fixed problem where not setting a footer caused the output document to be malformed.

Version 9.7.1🔗︎

December 3, 2021

New Features🔗︎

CV/Resume & Job Parser🔗︎

Hebrew is now a supported language, and Israel is a supported locale.

Version 9.7.0🔗︎

November 18, 2021

Improvements🔗︎

Search & Match🔗︎

/searcher endpoint now returns results sorted by a relevancy score instead of simply revision date.Learn more.

Regions in Australia are now normalized during indexing and querying.

Bug Fixes🔗︎

Search & Match🔗︎

Fixed indexing documents where address contains multiple regions.

When an address is specified during geocoding, replace the address on the parsed document so the new address can be indexed along with the latitude and longitude. This can then be used for exact location filters.

Version 9.6.9🔗︎

November 5, 2021

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better LinkedIn parsing in multiple languages.

Eliminated some minor runtime bugs.

Improved location parsing.

Better Hungarian education parsing.

Version 9.6.8🔗︎

September 22, 2021

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Internal query and index optimizations.

Improved parsing of social media handles and URLs and monikers.

Better inference of Swedish date of birth from national identity number.

Improved robustness of PII scrubbing.

Improved LinkedIn parsing.

Reporting more accurate certifications, and fewer certification variants.

Improved Work History parsing for all languages, including date parsing.

Improved Russian language location parsing, sectioning, and Work History parsing.

Version 9.6.7🔗︎

August 9, 2021

Improvements🔗︎

Search & Match🔗︎

Internal query and index optimizations.

Document Converter🔗︎

Improved image extraction/processing for large images.

Fixed timeouts in some documents.

Version 9.6.6🔗︎

August 3, 2021

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Fixed several runtime bugs.

Version 9.6.5🔗︎

July 27, 2021

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better parsing for personal information.

Improved sectioning of resumes, and identification of more education sections.

Better reordering of multi-column resumes.

Better parsing of LinkedIn resumes.

More accurate work history parsing, primarily in English, Chinese, German, Dutch, and Italian.

More accurate contact information parsing, especially for India and Czech and Norwegian languages.

More accurate education parsing, especially in French and Dutch.

Improved parsing for Romanian language.

Added more known certifications and personal qualifications.

Document Converter🔗︎

Better detection/extraction for candidate images.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed runtime bug related to anonymizing certain resumes.

Version 9.6.3🔗︎

June 30, 2021

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better location parsing for Pakistan.

Improved LinkedIn parsing to support the latest LinkedIn PDF changes.

Recognize the following online education websites:

Udemy
Coursera
Pluralsight
Udacity
CodeAcademy
Skillshare
edX
DataCamp
Treehouse

Better error handling.

Now we return an HTTP Status Code 413 - Payload Too Large for resumes or jobs that are too large to be parsed. Previously, these files would return 422 - Unprocessable Entity. The file size limit is ~6MB, which becomes ~8MB after Base64-encoding has been done. Files larger than this are typically not resumes/jobs, or are scanned images of resumes/jobs which Textkernel does not support. Scanned images need to be OCR'ed prior to sending to Textkernel to parse.

Version 9.6.2🔗︎

June 25, 2021

Improvements🔗︎

Document Converter🔗︎

Better candidate image recognition to eliminate many false-positives.

CV/Resume & Job Parser🔗︎

Greatly enhanced support for traditional Chinese language.

Improved candidate name and qualification suffixes parsing.

LinkedIn parsing enhancements.

Better Australian locale parsing.

Improved school parsing for Netherlands and Australia.

Better parsing for Canada, Portugal, China, Hungary, and Czech Republic.

Improved Contact Info, Work History, and Education parsing across all languages.

Better parsing for US addresses with no street address or postal code.

Added support for the following social media usernames/handles in Contact Info:

WhatsApp
GitHub
Skype
Instagram
Stack Overflow
Signal
Reddit
Quora
ICQ
WeChat
QQ
Telegraph
Telegram
MeWe
Parler
Gab

Simplified and improved candidate summaries across all languages.

Improved support for multi-column resumes.

Improved Job Level categorization in position history output.

Bug Fixes🔗︎

Document Converter🔗︎

Replaced functionality in third-party converter that was causing a memory leak when processing some PDFs when OutputCandidateImage was set to true.

Sourcing🔗︎

Fixed a runtime bug when purchasing resumes through CareerBuilder.

Version 9.6.1🔗︎

March 25, 2021

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Much better parsing for two-column resumes, including resumes that are only partially two-column.

Improved detection/parsing of date ranges that are split into a vertical multi-line section within an otherwise horizontal text section.

Better contact info parsing. We now report the best-formatted phone number when multiple versions of the same number are present in the resume. We no longer report fax numbers because it is 2021, not 1991. We no longer report work phone numbers when there are several other non-work phone numbers.

Improved recognition of stacked person names (names broken into words on separate lines).

We no longer report 'Politics', which was only active for Chinese resumes because we feel it is immoral.

Better education parsing. This is the most significant advance in education parsing in several years. Fewer false cognate school entries, better degree names, and much better major names. Also all certificates (except in certain locales such as Australia) have been moved to the Certifications section since they are really not the same as degreed education.

Far better recognition/handling of job nesting (or not) by company name and dates.

Improved accuracy on all aspects of employment parsing: company name, job title, location, and dates.

Improved parsing for LinkedIn resumes, especially for some non-English languages.

Improved parsing for Indeed's new resume format, which include an Assessments section.

Skills now generally ouput as Proper-case rather than UPPERCASE.

Improved accuracy across all data types in all languages, but especially so in English, German, Italian, Dutch, French, Russian, Hungarian, Czech, and Swedish.

Document Converter🔗︎

Improved PDF Conversion.

Upgraded 3rd party software.

Bug Fixes🔗︎

CV/Resume & Job Parser🔗︎

Fixed various reported bugs.

Document Converter🔗︎

Fixed a bug that was causing some blank HTML to be output on some corrupted/no-text documents

Version 9.5.2🔗︎

December 15, 2020

Improvements🔗︎

Document Converter🔗︎

Improved HTML Conversion.

Bug Fixes🔗︎

Search & Match🔗︎

Minor runtime bug fixes.

Version 9.5.1🔗︎

October 9, 2020

New Features🔗︎

CV/Resume & Job Parser🔗︎

Added initial support for Indonesian culture.

Improvements🔗︎

CV/Resume Parser🔗︎

More accurate Contact Info, Education and Employment parsing.

Speed improvements.

Minor bug fixes.

Document Converter🔗︎

Improvements to selection of correct conversion strategy.

Version 9.5🔗︎

September 3, 2020

New Features🔗︎

Search & Match🔗︎

Scoring and Default Sorting🔗︎

We have developed a new score (SovScore) that is derived from the combination of the WeightedScore and the Reverse Compatibility Score (RCS). Blending these two scores together using a proprietary algorithm allows us to provide a single score that blends results from both directional perspectives of the match. This new SovScore results in better matches rising to the top of the result set, while matches that score highly only from a single directional perspective will be scored lower.

All match results returned by the API, and in Textkernel’s own Search & Match UI, are sorted in descending order by the SovScore property. Please note that you are always free to use your own custom sorting or postprocessing of the returned results before displaying them in your own UI.

Transparency🔗︎

In an effort to continually increase the industry-leading transparency of our Search & Match engine, we have added new metadata to the match response. We added an EnrichedData property for each direction of the match. This object contains (a) the terms that were found, (b) the terms that were not found, and (c) explanatory messages (in English) about each data point. For example, in the skills object you might see that the skill 'java' was found, and in the corresponding explanation it might state that while 'java' was found in the document as a skill, it wasn't used by the candidate recently.

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Improved coverage for certain CAD/CAM skills.

Version 9.4.10🔗︎

August 26, 2020

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Better nesting of jobs based on employer names.

Better phone number parsing (eliminated some false cognates).

Better work history parsing in Dutch, Russian, Italian, French, and Greek.

Improved coverage of the cloud computing sub taxonomy.

Document Converter🔗︎

Can convert more documents to HTML, whereas some conversions previously failed.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Version 9.4.9🔗︎

August 6, 2020

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Improved rejection of false positive addresses and phone numbers.

Substantial improvements to skills taxonomy weights (expressed in output as percentages).

Better parsing for Dutch, Hungarian, and Croation languages.

Better accuracy in all sections.

9% faster.

Document Converter🔗︎

Better conversion to HTML and PDF.

Some files that could not be converted to HTML or PDF because of internal errors in the document can now be converted to HTML or PDF.

Version 9.4.8🔗︎

June 23, 2020

Improvements🔗︎

CV/Resume & Job Parser🔗︎

Remove some spurious HTML tags in plain text before parsing.

Better certifications parsing.

Better location parsing in all languages.

More accurate education parsing in all languages.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Minor runtime bug fixes.

Version 9.4.7🔗︎

May 29, 2020

Bug Fixes🔗︎

Job Parser🔗︎

Fixed a language identification bug when parsing non-English jobs.

Version 9.4.6🔗︎

May 22, 2020

Bug Fixes🔗︎

Document Converter🔗︎

Fixed a rare HTML conversion issue.

Search & Match🔗︎

Fixed various request validation bugs for malformed requests.

Version 9.4.5🔗︎

May 13, 2020

Improvements🔗︎

CV/Resume Parser🔗︎

Improved LinkedIn profile parsing.

Version 9.4.3🔗︎

May 1, 2020

Improvements🔗︎

CV/Resume Parser🔗︎

Better Hungarian parsing.

Better parsing of certifications.

Better parsing of educational majors.

Fewer false positives on educational GPAs.

Generally more accurate.

Fixed some minor runtime bugs.

Search & Match🔗︎

Improved the handling of punctuation in full-text searching.

Bug Fixes🔗︎

Search & Match🔗︎

Fixed edge case for handling bad input to CategoryWeights.

Version 9.4.2🔗︎

March 17, 2020

Improvements🔗︎

CV/Resume Parser🔗︎

Data improvements.

Document Converter🔗︎

Improved conversion of PDF to HTML.

More accurate detection of LinkedIn PDFs.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed two runtime bugs.

Version 9.4.1🔗︎

March 11, 2020

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed several runtime bugs.

Job Parser🔗︎

Additional runtime bug fixes.

Version 9.4.0🔗︎

March 4, 2020

Improvements🔗︎

CV/Resume Parser🔗︎

Better LinkedIn parsing.

World's fastest resume parsing is now substantially faster.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed several runtime bugs.

Version 9.3.10🔗︎

February 14, 2020

Improvements🔗︎

CV/Resume Parser🔗︎

Better LinkedIn parsing.

Better email parsing.

Better Chinese parsing.

Improved job title parsing for Bulgarian, Estonian, Finnish, Croatian, Hungarian, Lithuanian, Latvian, Polish, Romanian, Slovak, and Slovenian.

Improved education parsing for Indonesian school names.

Improved handling of company names which are actually URLs.

Improved handling of invalid dates such as February 29, 2007 (2007 was not a leap year).

Added a new configuration option to output all Metadata in English regardless of the resume language. Set "OutputFormat.AllSummariesInEnglish = true;" in the configuration string.

Job Parser🔗︎

Improved parsing for years of experience requirements.

Document Converter🔗︎

Better PDF Conversions.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed several runtime bugs.

Fixed bug that caused some jobs to be duplicated when reported on some German CVs.

Version 9.3.9🔗︎

November 21, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Better Work History parsing.

Improved LinkedIn profile parsing.

Document Converter🔗︎

Performance optimizations.

Version 9.3.8🔗︎

November 19, 2019

Improvements🔗︎

SaaS Services🔗︎

System performance optimization.

Version 9.3.7🔗︎

November 8, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Better Swedish work history parsing.

Better parsing for Educational History in all languages.

Improved skills taxonomy.

Improved LinkedIn profile parsing for location and street level address information.

Improved LinkedIn profile URL parsing.

Improved Candidate Name suffix parsing.

Improved recognition of some obscure date formats, primarily for European resumes in a language other than English.

Improved Chinese language parsing accuracy.

Improved accuracy of school names and degrees.

Document Converter🔗︎

Improved performance for converting documents to HTML.

Reduced timeouts on Excel documents.

Improved LinkedIn profile text conversion for the latest version of LinkedIn profiles.

SaaS Services🔗︎

Improved performance for all SaaS Service endpoints.

Version 9.3.6🔗︎

September 30, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Improved LinkedIn profile parsing.

Added new skills.

Improved Employment parsing for Swedish resumes.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed a runtime bug.

Version 9.3.5🔗︎

September 26, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Added parsing support for UK academic qualifications such as GCSE, BTEC, NVQ, and DipCG.

Search & Match🔗︎

Enhanced the robustness of the matching engine.

Added support for non-integer taxonomy ids for Match by Criteria requests.

Version 9.3.4🔗︎

September 12, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Better logic for two-character skills to prevent reporting erroneous data as a skill.

More accurate Company Name parsing.

Better Italian Employment History parsing.

More accurate education parsing.

Up to 20x faster on some resumes, and about 9% faster overall.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed several runtime bugs that were tested to occur about once every million documents.

Version 9.3.3🔗︎

August 28, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Improved parsing of LinkedIn profiles in Spanish.

Improved the Experience Summary output when no dominant taxonomy was found.

Improved phone number parsing.

Document Converter🔗︎

Improved detection and conversion of reversed text in poorly constructed PDF documents.

Search & Match🔗︎

Diacritics (e.g. é) can now be used in index names and document IDs.

Version 9.3.2🔗︎

August 14, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Major improvements to Employment parsing by changing the logic used for nesting of jobs by date ranges. The logic was improved all around, meaning that more jobs that should be nested do get nested, and more jobs that should not be nested don't get nested.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Fixed several runtime bugs that were tested to occur about once every million documents.

Document Converter🔗︎

Fixed a condition that would report a file as corrupt rather than having too-short-to-be-believed lines.

Version 9.3🔗︎

July 29, 2019

New Features🔗︎

CV/Resume Parser🔗︎

Added full support for the following languages/locales:

Bulgarian
Croatian
Estonian
Finnish
Hungarian (was previously partially supported)
Latvian
Lithuanian
Polish
Romanian
Slovak
Slovenian

Job Parser🔗︎

Added full support for the following languages/locales:

Bulgarian
Croatian
Estonian
Finnish
Hungarian (was previously partially supported)
Latvian
Lithuanian
Polish
Romanian
Slovak
Slovenian

Search & Match🔗︎

Added full support for the following languages/locales:

Bulgarian
Croatian
Estonian
Finnish
Hungarian (was previously partially supported)
Latvian
Lithuanian
Polish
Romanian
Slovak
Slovenian

Bug Fixes🔗︎

Search & Match🔗︎

Fixed a bug where setting MatchCriteria.MonthsManagementExperience = null would cause an error.

Version 9.2🔗︎

July 18, 2019

Improvements🔗︎

CV/Resume Parser🔗︎

Accuracy improvements in all languages.

More accurate contact info parsing. Better phone number type classification.

Job Parser🔗︎

Parsing throughput has increased over 50%.

Document Converter🔗︎

Better detection and correction of reversed text.

More useful conversion of LinkedIn pdf profiles in the varied two-column formats.

Faster conversion of PDF to HTML.

Better trimming of leading whitespace preceding first non-whitespace character in converted text.

When documents take too long to convert, we now return converted text in many more instances, whereas before we returned no text.

Improved messaging for possible conversion errors.

New Output Validity: ovLinesSeemTooShort.

Added messages for output validity warnings and errors.

Upgraded third party DLL versions:

Aspose Words 19.4
Aspose PDF 19.4
dtSearch 7.93.8596.18093

Search & Match🔗︎

We significantly improved normalization of school names behind the scenes, so that searches on school names will return more comprehensive results. We also stopped indexing high school names to reduces false positives.

Bug Fixes🔗︎

CV/Resume Parser🔗︎

Improved non-English Textkernel-generated candidate summaries.

Fixed bug that caused some jobs to be duplicated when reported on some German CVs.

Fixed an issue where skills and normalization data could be cached for longer than 24 hours.

Document Converter🔗︎

Corrected situation where some non-binary data was being detected as binary data and being reported as ovProbableGarbageInText.

Search & Match🔗︎

Fixed an issue applying an upper bound of years experience for a skill.

Fixed an issue filtering multiple document languages in the same transaction.

Fixed an issue where requests would fail when searching or matching across multiple indexes that have documents with the same document id in the results.

Fixed an issue where match results would sometimes be out of order when sorted by the score.

Version 9.1.2🔗︎

November 13, 2018

Improvements🔗︎

CV/Resume Parser🔗︎

7% faster.

Parses all LinkedIn past and present versions extremely accurately.

Better Swedish date parsing.

More accurate employment parsing.

More accurate education parsing.

Improved resume sectioning.

Document Converter🔗︎

Better PDF conversions.

Version 9.1.1🔗︎

November 8, 2018

Improvements🔗︎

CV/Resume Parser🔗︎

Fixed management level output for resumes with no current employment.

Search & Match🔗︎

Improve Bimetric Scoring in cases where no second-best taxonomy is found.

Better comparison algorithm for job titles that contain prepositions.

Improved languages matching algorithm.

Version 9.1🔗︎

October 20, 2018

Improvements🔗︎

CV/Resume Parser🔗︎

Greatly improved parsing of gradepoint averages in Education.

Greatly reduced the number of spurious trailing work history jobs or educational schools.

Thousands of improvements to internal data lists.

Vastly improved LinkedIn parsing. We are now able to capture the hidden LinkedIn urls, and ignore the broken partial LinkedIn urls.

Degrees which are just certifications and not intended to be high school-or-higher degrees are now not output in Education, but rather, are output in Certifications.

Better parsing of school names. Fewer school names with City names hanging on the end (sometimes they need to be left that way; other times they need to be stripped – we do both better now).

Better parsing of Russian, Italian, and Norwegian schools and degrees.

Far more accurate nesting of PositionHistory nodes within EmployerOrg nodes: specifically, far fewer wrongful nesting events, and a few more correct nesting events.

We restored and improved the parsing accuracy for BOTH past and present LinkedIn resumes in all known formats.

Improved Company Name and Position Title accuracy by several percentage points. Improved the ability to distinguish between ambiguous elements.

Document Converter🔗︎

Better removal of page numbers.

Vastly improved LinkedIn conversions. Conversion to single column format now happens in correct order. Page markers are properly removed. Broken lines are re-connected.

Have real formatted HTML output available from PDFs now.

Improved HTML-to-text conversions. HTML should not contain tabs except within <pre> tags, but some HTML wrongfully does. In the past, these tabs were converted to a single space; now, we convert them to multiple spaces. This ends up allowing the Parser to “see” many more section headers that in the past were invisible because they collided with nearby words.

Bug Fixes🔗︎

Fixed a bug in the ReservedData section output that would cause an error in scrubbing PII.

We were eliminating some valid URLs. We fixed that so that we now report more URLs.

Version 9.0.2🔗︎

September 7, 2018

Improvements🔗︎

Better PDF conversions to fix some things intentionally broken by LinkedIn.

Version 9.0.0🔗︎

May 18, 2018

Upgrade Path🔗︎

If you are upgrading from version 8.0 or later, switching to version 9 is as simple as changing the url of the service from v8 to v9. No other changes needed, typically.

If you are upgrading from version 7.5 or earlier, this version isn't compatible with version 7.5. To upgrade to version 9.0, we recommend the following approach:

If you're using a parser configuration string, regenerate your string in the new human-readable Name=Value pair format.
Parse the Sample.doc file (as well as some of your own documents) in the current version you use, and with 9.0 using our Demo Application and save those results to disk.
Use a document comparison tool to evaluate the differences, specifically the new fields. There is a lot of new metadata provided that could be of high value to integrate in your application. These new fields are detailed below in the New Features section. For a document comparison tool, we really like Beyond Compare.
Remap your API calls to the new 9.0 methods as described in the API Documentation, make the desired changes to your implementation to leverage the new metadata, change the URL to point to version 9.0, and enjoy.

New Features🔗︎

Added an endpoint to scrub the Personally Identifiable Information from a Resume/CV. More information can be found in the REST API documentation.

Improvements🔗︎

CV/Resume Parser🔗︎

Improved the skills taxonomies for all languages. We added a new taxonomy/Subtaxonomy for all languages: "No dominant taxonomy → Not enough data". When we cannot determine the taxonomy with confidence because so few (or no) skills were found, we output "No dominant taxonomy → Not enough data".

Improved accuracy on Work History and Education.

Improved sectioning of resumes.

Overall accuracy is up about 3 absolute percentage points, with 99% of the previous speed. Textkernel parsing speed is typically at least 5x faster than our nearest competitor’s speed, and we produce about 1/3 to 1/10 of the mistakes as our nearest competitor.

Search & Match Engine🔗︎

Improved the handling of management level queries in Matching when there was no management level data in the source document.

Breaking Changes🔗︎

Skills🔗︎

We deprecated the SkillsStyle property because we now have a single canonical way and place to output skills.

Skills are now output only in the resume's UserArea, or job's SkillsTaxonomyOutput. The output is extremely easy to read and understand from both a human and programmatic standpoint. The output taxonomies are sorted in descending order of importance, and skills are alphabetical within the subtaxonomies, and child skills are nested within parent skills.

Also, importantly, we now use the English skills list for non-English skills parsing in addition to the detected language's built-in skills list. This will generally result in more skills being found, with very few false cognates.

DO NOT use/rely on the skill Ids that are output. We reserve the right to modify skill names and to preserve the skill Id when we do so. In some cases, we append a language code to skill Ids so that we can output them alongside another translation of that skill with the same Id. If you are relying on skill Ids, stop!

NOTE FOR CUSTOM SKILLS LISTS: When developing your custom skills lists, you must avoid using ANY Textkernel taxonomy or skill Ids. The only way to be certain of that is to prepend or append an alphabetical character to your Ids if they are only integers.

Other🔗︎

We deprecated the ParserSettings.OutputFormat.ReportAllCompanyNamesAndPositionTitlesRegardless and ParserSettings.OutputFormat.ContactMethod properties.

We made these properties read-only:

ParserSettings.OutputFormat.XmlFormat
ParserSettings.OutputFormat.MinimumCompanyNameProbability
ParserSettings.OutputFormat.MinimumPositionTitleProbability

We moved the Bimetric Score endpoint from /bimetricanalyzer to /scorer/bimetric.

Bug Fixes🔗︎

Fixed an uncommon issue in our JSON output where some arrays were output as objects when they ony had a single item.

Fixed an issue where the OutputFormat.NormalizeRegions parser setting was being ignored.