See How My Reps Align

Did your representatives represent the stances you declared? This is the accountability spine of PolitiClaw: everything else (bills, ballots, monitoring, outreach) feeds into or off of the answer.

By default the workflow is deterministic — there is no LLM judgment about how a rep "really feels" about an issue. You declare stances with weights, PolitiClaw matches those stances to bills by subject and title, you record agree/disagree signals on specific bills, and your rep's federal roll-call votes (House via api.congress.gov, Senate via voteview.com) are counted for or against your direction. Scores below the confidence floor surface as insufficient data, never as a false percentage.

The optional Auto-Rated Direction feature lets a host LLM fill in implied direction for bills you haven't manually rated, with grounded-quote guardrails and explicit per-score disclosure. It defaults to off; turn it on only after reading the trade-offs.

Rep coverage is US House and US Senate only; state-legislative and local offices are not resolved — see current coverage.

Default tools

• politiclaw_get_my_reps — resolve your current federal delegation from your saved address.
• politiclaw_score_representative — per-issue alignment for one rep, with cited aligned/conflicted bills.
• politiclaw_rep_report — recompute the whole delegation in one pass; intended for the monthly accountability digest.

Chain them in order: get_my_reps → score_representative (one rep at a time while you're exploring) → rep_report (when you want the combined document).

Setup prerequisites

Rep accountability requires declared issue stances and at least one recorded stance signal on a relevant bill. If scores come back as insufficient data for every issue:

1. Review your declared stances with politiclaw_issue_stances and add more if the set is thin.
2. Record agree/disagree signals on tracked bills from the dashboard's quick-vote section or the politiclaw_record_stance_signal tool. With auto-direction off (the default) direction comes only from your explicit signals; with Auto-Rated Direction on, the host LLM can supply direction for unrated bills under modes you choose.
3. If roll-call data is missing, run politiclaw_ingest_votes first (defaults to both chambers).

What's counted — and what isn't

• In scope: federal House (api.congress.gov) and Senate (voteview.com) roll-call votes, procedural motions excluded by default, bills your issue stances match by subject/title.
• Insufficient data: any issue whose confidence falls below the floor; reps with no ingested votes yet; any bill you haven't recorded a direction signal on (and that the optional auto-direction feature hasn't filled in, if enabled).
• Out of scope: state, local, and gubernatorial accountability. PolitiClaw does not claim coverage it does not have — if a rep isn't in the federal House or you haven't wired up the state sources, the report will say so rather than guess.

Focused follow-ups

• politiclaw_election_brief when the accountability findings should fold into ballot context.
• politiclaw_draft_outreach when the next step is outreach grounded in a cited conflict.
• politiclaw_review_auto_ratings (auto-direction users only) for bills the AI flagged as borderline / mixed / unclassifiable so you can apply judgment.

In the accountability loop

This page is the operational entry point. The other three sides of the loop:

• How PolitiClaw Holds Representatives Accountable — the full loop with the diagram, the dissenting-view rule, and source-tier discipline.
• Track Bills and Votes — the evidence side: how bills and roll-call votes get into the system before scoring runs.
• Examples of Good Alerts — what the periodic rep report and weekly digest actually look like.