pipelineclusteringweb-app
Smarter keyword discovery, faster clustering, and a new web app workflow
Shipped32 commits+22362−1772
- Keyword discovery now completes the full loop — Supr Keywords now runs the complete related → ideas → suggestions flow, so keyword discovery goes further with less manual work.
- Clustering is now faster and works on much larger datasets — We added automatic clustering selection for small and large jobs, plus a faster K-means path that handles thousands of keywords more reliably.
- The web app now covers the full job workflow — Users can create jobs, review results in dedicated tabs, export keywords, and manage jobs from a cleaner dashboard experience.
- pipelineAdded keyword suggestions expansion so discovery now continues from related keywords and ideas into suggestions.
- clusteringAdded semantic clustering with cluster summaries and pillar keyword selection to group related keywords more meaningfully.
- clusteringAdded automatic clustering selection, using simpler clustering for smaller jobs and K-means for larger ones.
- web-appAdded the web app job workflow, including job creation, results tabs, and job deletion from the dashboard.
- exportsAdded CSV export for keyword results, including cluster labels and smarter filenames.
- pipelineMade keyword exclusion smarter by focusing on the highest-opportunity keywords instead of removing results at random.
- dashboardImproved job browsing with clearer domain names, locale codes, status badges, and automatic sidebar refreshes.
- dashboardAdded search, sorting, and pagination to the Keywords and Clusters views for easier exploration of large jobs.
- dashboardAdded monthly volume charts and friendlier language and location names to make results easier to read.
- pipelineImproved progress messages so long-running jobs show clearer, more user-friendly updates.
- web-appFixed language and location dropdowns so they now load and select the right options correctly.
- clusteringFixed clustering issues that could place all keywords into one group or leave matching keywords out.
- exportsFixed CSV export problems so large downloads label clusters correctly and handle special characters safely.
- dashboardFixed job status handling so failed jobs no longer get stuck in running or processing states.
- web-appFixed several build and runtime issues in the web app to make local development and deployment more reliable.
- apiAdded versioned job endpoints for creating jobs, listing jobs, viewing job details, and deleting jobs.
- apiAdded results endpoints for personas, keywords, clusters, and cluster keywords, with filtering and pagination support.
- apiAdded language and location endpoints for the job form, plus support for downloading keyword data as CSV.
- apiImproved cache control endpoints so teams can clear data by domain or by specific locale when needed.
- clusteringReduced clustering time dramatically for large jobs by using a faster K-means path and batching embeddings.
- pipelineLowered keyword discovery costs by limiting suggestion requests and applying smarter API-level filters.
- web-appImproved form speed by loading language and location lists from built-in data instead of runtime external requests.
- dashboardReduced unnecessary polling and API usage across the dashboard and pipeline.
- docsAdded onboarding guides for the web app, pipeline, database, and shared utilities.
- docsUpdated the web app PRD, roadmap, and changelog notes to reflect the new dashboard and clustering work.
- docsAdded planning and onboarding docs for future authentication, credits, and billing work.
- apiLanguage and location APIs now return built-in data only, so any integrations expecting live external lookups should be updated.