Skip to content
Follow yggy on LinkedInFollow @yggy_tech on X

September 2026

Each day starts with a short summary, followed by the details.

Review → Ontology now collects identity and vocabulary proposals in one inbox that you work through item by item. Data Explorer and the projects list also cope better when something goes wrong.

  • Feature Review → Ontology has one inbox for identity and vocabulary proposals. You decide on them one at a time.
  • Improvement Ontology search stays focused on what you type, and the vocabulary tabs are aligned.
  • Fix Data Explorer search is more reliable, lets you retry when a request fails, and shows the right icons.
  • Improvement The projects list shows who has access to each project and lets you retry when loading fails.
  • Fix Graph totals still show when the server is slow to answer.
  • Fix Wording and breadcrumb mistakes in the interface are corrected.
  • Chore Extraction is more careful on its most accurate setting, and we prepared exporting your ontology in standard formats.

yggy now has a notes app for personal accounts, with Free and Plus plans. Projects gain Cases, a workbook for investigations such as 8D reports, and you can import from Slack and Notion, ask chat about a specific source, and connect Codex and Copilot to yggy.

  • Feature A new notes app for personal accounts: capture a note and yggy picks out the people, companies and topics in it.
  • Feature The notes app has a Free and a Plus plan. You can upgrade from Free to Plus with a card.
  • Feature Personal accounts can connect AI assistants to their notes. Free includes note tools, and Plus adds related-topic lookups. Organization access stays separate.
  • Feature Cases replace Investigations. Each project has a case workspace, for example for 8D reports, laid out as a canvas that you fill in like a workbook.
  • Feature Import Slack and Notion content straight from the picker.
  • Improvement Sources open in a browser that works like Google Drive, with a “show more” link, and you can sort them by last modified.
  • Improvement Google Drive import now handles Markdown and PowerPoint files and shows a clear message when an import fails.
  • Feature From a source’s page you can ask chat about that document. Your question shows as a chip, and the source opens reliably from chat.
  • Fix A document in which no facts are found now finishes with a clear message instead of trying again and again.
  • Improvement When a document stops because you reached a usage limit, retrying shows the reason and a link to billing.
  • Feature Sections of a document’s content have copy buttons.
  • Improvement The built-in ontology covers more financial terms.
  • Feature You can connect Codex and Copilot to yggy. A connect page shows the server address, ready-to-paste setup for each tool and your organization key.
  • Improvement Project cards are tighter, easier to scan and show placeholders while they load. Project lists load page by page and put the project you viewed last first.
  • Improvement A shared back button across pages, and the upload dialog offers to connect a data source.
  • Feature Our team can set your plan and credit cap for you.
  • Improvement All emails from yggy share one consistent look.
  • Improvement Usernames can be as short as 2 characters.
  • Fix Pages appear in your language right away, without switching language after loading.
  • Improvement Full-screen loading shows a new animated spinner, and inside the app, loading shows placeholders.
  • Improvement Documents, Projects, Analytics, Ontology and Review load faster, the login page is lighter, and Analytics keeps its graph totals when the server is busy.
  • Fix Review lists merges per project, and the list of possible duplicates no longer fails or stalls.
  • Fix Graph Explorer no longer keeps showing a loading placeholder after the project’s documents have loaded, and Data Explorer finds your project from your membership.
  • Fix Chat tools reconnect after a dropped connection, and an empty search no longer fails.
  • Chore More reliable text recognition and relation extraction, product analytics, tighter checks around sign-in consent and project access, and many stability and speed fixes.

Review is easier to work through: conflicting facts get their own cards and likely duplicates are listed by how closely they match. Uploads now tell you why a document failed or stalled, and setting up AI assistants is clearer.

  • Improvement Review shows conflicting facts on their own cards, also when two values for the same fact tie.
  • Improvement Review lists likely duplicates by how closely they match, with a badge on each.
  • Improvement Uploads show why a document failed or stalled.
  • Improvement Setting up AI assistants groups tools by type and can be done for the whole organization. Each tool shows how many connections it has, with a page of its own, and the whole tile is clickable.
  • Improvement opencode and Gemini show their official logos, and connections from unknown tools appear on the Other tile.
  • Improvement The consent screen is organized into collapsible sections.
  • Feature Six new data sources: Slack, Notion, GitHub, GitLab, Jira and Confluence.
  • Improvement Search also finds codes, names and numbers captured from your documents.
  • Improvement Serial, customer and lot numbers are recognized as identifiers by default, and the same number in several files leads to one approved entry.
  • Fix Retired project terms are no longer offered.
  • Improvement yggy asks you to confirm destructive actions, and the number of seats on your plan follows the number of members in your organization.
  • Fix The screen no longer stays locked after you create a folder or add a source, and document stat cards no longer overlap on narrow screens.
  • Fix Graph analytics and recent changes load faster on large organizations and no longer fail there.
  • Chore Rarely used facts can move to cheaper storage, graph updates and retries are more reliable, identifiers are handled more consistently, and our services report their health in a standard way.

You can now connect six more data sources, and identity review keeps moving in projects without an identity policy.

  • Feature Six new data sources: Slack, Notion, GitHub, GitLab, Jira and Confluence. Connect one to browse and import its content.
  • Fix Identity review no longer stalls when a project has no identity policy.
  • Improvement Connections from an unknown AI tool now appear on the Other tile during setup.

Search now reaches into the codes, names and numbers captured from your documents, and big organizations can run large graph queries without errors.

  • Feature Search now covers codes, names and numbers captured from your documents.
  • Fix Large graph queries no longer fail on very large organizations.
  • Fix Searching stored values works again when your filter contains special characters.

Retired terms stay out of your way, key values from your documents are kept as evidence, and the AI assistant setup is better organized.

  • Fix Retired project terms now stay retired: extraction and review no longer offer them.
  • Improvement Key facts such as codes, names and numbers are now kept as evidence in their own right.
  • Improvement The AI assistant setup page groups tools by type and supports setup for the whole organization.
  • Improvement Connection lists show how many connections each tool has, with a page for each tool.

Extraction is more dependable: your graph is updated completely, and facts your ontology does not cover wait for your review.

  • Fix Graph updates from an extraction now complete fully. Facts that the project’s ontology does not cover wait for review instead of being written.
  • Fix Identifiers of merged entries are now kept consistently.

Product material numbers are now handled the same way in every document.

  • Fix Product material numbers are now recognized the same way in every document.

Structured JSON files keep more of their data, and finished documents no longer look stuck.

  • Fix JSON files that need new property names still produce facts for review instead of finishing with an empty graph.
  • Fix JSON files now keep the codes and names that were already mapped, also when the record type is new. A source id without an approved identifier space no longer drops the whole file.
  • Fix Finished documents no longer look stuck on vocabulary review.
  • Fix The From, To and Subject fields of emails are no longer treated as topics in your graph.