Merged
Conversation
…ity fixes, visible errors, cost tracking, and shareable runs Ships ten new user-facing capabilities plus two critical fixes: • New Blind Jury engine (parallel one-shot responses with zero cross-visibility + optional Judge synthesis) alongside the existing CVP. Toggleable from the Protocol panel. • CVP now defaults to blind Round 1 (no first-mover contamination) and randomised participant order in later rounds; both toggleable. Adds convergence-based early stopping and a non-voting Judge that preserves minority views. • Failed provider calls now show clear errors instead of a bare "…". All providers' OpenAI-compat shims now actually work (pinned to the real chat-completions endpoint). • Live cost meter (tokens + USD estimate) and per-participant confidence trajectory sparkline. • Disagreement panel that groups conflicts by round and lets you jump straight to the source messages. • Prompt library with eight curated presets (Strategy / Engineering / Ethics / Science). • One-click session export (Markdown/JSON) + permalink sharing via compressed #rt=… hash. Anyone opening the link gets a read-only replay with exit banner — no re-running calls. • Floating right-side panel on large screens (CostMeter + ConfidenceTrajectory + DisagreementPanel + MessageFlow) that stays visible; falls back to sidebar on smaller screens. • Full ConfigPanel for every new toggle plus judge model selection. All changes are live in the demo, with the README fully rewritten to reflect the new capabilities, engines, and sharing flow.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Ships ten new user-facing capabilities plus two critical fixes:
• New Blind Jury engine (parallel one-shot responses with zero cross-visibility + optional Judge synthesis) alongside the existing CVP. Toggleable from the Protocol panel. • CVP now defaults to blind Round 1 (no first-mover contamination) and randomised participant order in later rounds; both toggleable. Adds convergence-based early stopping and a non-voting Judge that preserves minority views. • Failed provider calls now show clear errors instead of a bare "…". All providers' OpenAI-compat shims now actually work (pinned to the real chat-completions endpoint). • Live cost meter (tokens + USD estimate) and per-participant confidence trajectory sparkline. • Disagreement panel that groups conflicts by round and lets you jump straight to the source messages. • Prompt library with eight curated presets (Strategy / Engineering / Ethics / Science). • One-click session export (Markdown/JSON) + permalink sharing via compressed #rt=… hash. Anyone opening the link gets a read-only replay with exit banner — no re-running calls. • Floating right-side panel on large screens (CostMeter + ConfidenceTrajectory + DisagreementPanel + MessageFlow) that stays visible; falls back to sidebar on smaller screens. • Full ConfigPanel for every new toggle plus judge model selection.
All changes are live in the demo, with the README fully rewritten to reflect the new capabilities, engines, and sharing flow.