Discord export
Forum thread messages, authors, dates, reactions, attachments, and thread IDs were exported from the TDR Discord.
CONTEXT
Community feedback packaged for development review.
We exported structured feedback threads from Discord, analyzed them with two different AI models, and produced one assessment for each thread. Those assessments are plugged into this static site so developers can browse the feedback by review lane, ask, sentiment, source voices, and dev notes. The content and code are available at github.com/imajes/tdr-feedback.
How the assessments were produced
A compact chain from raw community discussion to reviewable site content.
Model pass A
One AI model extracted candidate asks, sentiment, representative voices, and defect evidence from each thread.
Model pass B
A second model cross-checked the synthesis, tightened wording, and helped separate bugs from opinions.
Markdown assessments
Each thread assessment remains in /assessments as the editable source of truth for the site.
Static site
The generator turns the assessments into Cloudflare-ready HTML, CSS, JSON, sitemap, and headers.
Data export
The generated JSON file exposes the same site model used by the rendered pages.
/data.json
The export contains the review lanes, visible thread assessments, top priority asks, and aggregate site stats. It is useful for quick inspection, downstream analysis, or rebuilding the site in another presentation layer.
Open JSON exportbuckets[]- Review lanes with route IDs, labels, icons, descriptions, and included thread records.
threads[]- Published assessments with metrics, asks, bug findings, sentiment, source links, voices, and dev notes.
topAsks[]- Cross-thread priority asks with source route, category, score, and thread-level signal metrics.
stats- Total assessments, visible threads, participants, messages, screenshots, lanes, and build date.