Evidence Binder

• Lays out every evidence item in your Case File two ways — grouped by criterion, or grouped by the source document it came from — with no AI model call involved at all: it's a deterministic read of your existing Case File data, per Merito's "code before AI" rule.
• Assigns exhibit numbers automatically and flags real gaps as you go: missing source URLs, missing page references, missing issuer or writer's address on letters, evidence that backs several criteria, and archive status for anything saved to the Wayback Machine.
• Surfaces a "Consider mapping" section for strong evidence that isn't yet linked to any criterion — the kind of thing that's easy to add and then forget to actually use.

Screenshot placeholder: no live capture this pass — see the report's evidence table for why. Note that unlike this help series' AI tools, Evidence Binder has no "run" to wait for — the fixture case's own 5 real evidence items (per docs/DEMO_CASE.md) would render on this page immediately, with no analysis step, whenever a screenshot is captured.

Evidence Binder sits in the Position stage of the Case Journey (journeyStageMeta.ts), alongside Narrative Studio and Petition Studio — after you've gathered evidence, before you stress-test or draft around it.

• Before: Citation Context Analyzer and other evidence-gathering tools feed evidence directly into your Case File — the Binder is where that evidence actually gets organized and numbered, not where it originates.
• After: the exhibit numbers assigned here are what Denial/NOID Response and RFE Studio's objection cards reference when they link an objection to "the exhibit that answers it." Devil's Advocate and Debate Mode also cite specific exhibits from this same numbering when arguing your case.

• Input: nothing to type or upload — it reads directly from your existing Case File's evidence items.
• Case File prerequisite: yes, implicitly — the Binder only shows what's already in your Case File; add evidence via Case Canvas or another tool's "Add as evidence" action first.
• Typical time: as long as it takes to review your evidence and fix any flagged gaps — there's no run or wait involved, since nothing here calls a model.
• Credit cost: free to view (0 credits) — but it's a Pro-only tool, so a free-plan account can't open it at all.

• Exhibit N — the automatically assigned exhibit number for that evidence item, consistent with what Denial Response's and RFE Studio's objection cards reference.
• Evidence-strength badge — the same weak/moderate/strong badge used everywhere in Merito; "ungraded" appears if an item hasn't been scored yet.
• Criterion strength dot color — in the sticky nav, each criterion's dot reflects its aggregate evidence strength across all linked items (gap / weak / moderate / strong).
• Archive status — "archived" (saved to the Wayback Machine), "pending" (archive in progress), or "failed," shown for web-sourced evidence.
• Backs several criteria badge — marks evidence that supports two or more criteria, worth strengthening first because it lifts all of them.
• Missing-metadata warnings — separate badges for a missing screenshot source URL, missing page reference, missing issuer, or missing writer's address, each naming exactly what's absent.
• Source document removed — a distinct, clearly-explained bucket for evidence whose original uploaded file was later deleted; the evidence itself is kept, but flagged for review rather than silently mixed into "unsourced."
• Unsourced — evidence added directly rather than extracted from a file — expected and not itself a problem.
• Consider mapping — strong evidence with no criterion link yet.

• Evidence Binder is not on the Free plan — it's explicitly listed in plans.pro.features_json (evidence_binder), gated by plan even though it costs 0 credits to actually use once unlocked.
• A 7-day Pro trial gives you unlimited use of the Binder for the trial's full duration — there's no per-run credit cost eating into your 40-credit allotment for this specific tool.
• The trial checkout collects a card up front (a standard Stripe subscription checkout with a trial period, not a card-free trial), but cancellation is straightforward — cancel anytime from Settings before the trial ends and you won't be charged. Start your free trial to see your own case organized this way.

• Leaving strong evidence unmapped to any criterion. The "Consider mapping" section exists precisely because this is easy to do by accident — check it before you assume your binder is complete.
• Ignoring missing-issuer or missing-writer's-address warnings on letters. These are exactly the kind of gap a real reviewer notices; fix them via the inline edit action rather than leaving them flagged.
• Assuming a deleted source document also deletes its evidence. It doesn't — the evidence item survives in a clearly-labeled "source document removed" bucket, which still needs your attention (re-link a replacement source if one exists).
• Not checking archive status on web-sourced evidence. A "failed" or "pending" archive status means that evidence isn't yet safely preserved against link rot — worth re-triggering or following up on.
• Expecting the Binder to grade or generate evidence. It's a deterministic organizer, not an analysis tool — it shows you what strength scores already exist, it doesn't create new ones.