Civic Accountability & Intelligent Content Creation
User Manual — Version 1.0  |  March 2026

RhetoricalPoints is a civic accountability platform. It uses artificial intelligence to check political claims, analyze debates, and create accountability content — automatically. This manual walks you through every feature, step by step.

Copyright © 2026 Stephen R. Thompson / RhetoricalPoints LLC. All Rights Reserved.
stephen@rhetoricalpoints.com
Scroll to start

What is RhetoricalPoints? It is a web app that checks whether political claims are true. It uses several AI models at once, then compares their answers. You get a verdict, a score, and the sources that back it up. You can check a single claim, a full speech, or an entire debate. The platform also creates fact-checking videos automatically — no editing skills needed. Everything runs in your browser. No software to install.

Current App Layout

The product now uses cleaner guided entry pages in several places. If a section in this manual sounds broader than what you see on screen, start from the focused route first and then follow the step-by-step flow.

Live Debate Starts from separate mode pages and then moves into dedicated setup steps.
Content Studio Video analysis, scripts, debate analysis, settings, and pipelines now live on separate creator pages.
Speech & Debate Trainer Practice lanes now begin from cleaner route pages before opening the live practice room.
Leaderboard The public placeholder page is live, but the full ranking system is still being held back intentionally.
Table of Contents
Part One
Fact-Checking Tools

These tools check claims for accuracy. You can check a single statement, a full speech, or a live debate. Every check runs across multiple AI models at once.

🔍
Feature 01
Truth Engine — Claim Checker
Paste any claim. Multiple AI models check it at the same time. You get a single verdict, a score, and the sources that back it up.
Who it's for Anyone who wants to fact-check a specific statement
Access level All Users
Time to result 10 – 30 seconds

How to use it

  1. 1 Go to Truth Engine in the top navigation bar.
  2. 2 Paste or type the claim you want to check into the text box. Use the exact words from the politician or speaker.
  3. 3 Click Verify Claim.
  4. 4 Wait for the results. The system runs up to five AI models in parallel. You will see a progress indicator for each one.
  5. 5 Read the consensus verdict at the top. It summarizes what the majority of AI models found.
  6. 6 Scroll down to see each model's individual verdict and reasoning.
  7. 7 Check the Web Sources section at the bottom. Click any link to read the original source.

Possible verdicts

True
False
Misleading
Partially True
Tip: The Integrity Score runs from 0 to 100. A score of 80 or above means the claim is well supported. Below 40 means it is likely false or misleading. The Consensus % shows how many AI models agreed on that verdict.
What you get

A verdict (True / False / Misleading / Partially True), an Integrity Score from 0 to 100, a Consensus % showing model agreement, a per-model breakdown with reasoning, and live web sources with direct links.

📄
Feature 02
Speech / Transcript Checker
Paste an entire speech or transcript. The AI finds every factual claim inside it. Then it verifies each one automatically.
Who it's for Journalists, researchers, and engaged voters
Access level All Users
Time to result 1 – 5 minutes (depends on speech length)

How to use it

  1. 1 Go to Content Studio, then open the Analyze Video path. That route now handles transcript-style checking as well as video input.
  2. 2 Paste the full text of the speech, statement, or press release.
  3. 3 Add the speaker's name and date if you know them. This helps the AI search for context.
  4. 4 Click the main analyze button to send the transcript into the Truth Engine workflow.
  5. 5 The AI reads the text and pulls out every verifiable claim. You will see them appear one by one as they are extracted.
  6. 6 The Truth Engine checks each claim. Watch the progress bar fill as each claim gets a verdict.
  7. 7 Review the Overall Integrity Score at the top of the results. This is an average across all verified claims.
  8. 8 Click any claim in the list to expand its full breakdown and sources.
  9. 9 Click Download Report to save the full analysis as a PDF.
What you get

An Overall Integrity Score for the entire speech. A list of every factual claim extracted from the text. A per-claim verdict with sources. A downloadable PDF report you can share or cite.

⚖️
Feature 03
Debate Analyzer
Paste a debate transcript. Get fact-check scores for every speaker. See which claims were flagged and why.
Who it's for Anyone analyzing political debates or panel discussions
Access level All Users
Modes Two-Person Debate  |  Panel Discussion

How to use it

  1. 1 Go to Debate Analyzer in the navigation bar. Start from the clean mode page first.
  2. 2 Pick a mode: Two-Person Debate for head-to-head formats, or Panel Discussion for group conversations.
  3. 3 Paste the debate transcript into the text area. Make sure each speaker's lines start with their name, like:
    SPEAKER NAME: their words here
  4. 4 Add the speaker names if prompted. The system can detect them automatically.
  5. 5 Click Analyze Debate.
  6. 6 Review the per-speaker integrity scores at the top. Each speaker gets their own score.
  7. 7 Scroll down to see flagged claims sorted by speaker. Each flagged claim shows its verdict and source links.
  8. 8 Read the Debate Summary at the bottom. The AI writes a short neutral overview of the key accuracy issues.
Tip: You can get debate transcripts free from C-SPAN, Rev.com, or the Congress.gov record. Copy and paste the full text — the longer the transcript, the more useful the analysis.
What you get

Per-speaker integrity scores. A list of flagged claims with verdicts and sources. A neutral AI-written debate summary. Exportable results for sharing or publishing.

🎙️
Feature 04
Live Debate
Run real-time fact-checking during a live debate. Each speaker submits statements. The Truth Engine checks them on the spot. Scores update live in the sidebar.
Who it's for Event organizers, debate moderators, classrooms
Access level All Users
Participants 2 or more speakers

How to use it

  1. 1 Go to Live Debate in the navigation bar.
  2. 2 Choose the type of live debate you want to run: Live Debate, Bot vs Bot, or Surrounded.
  3. 3 Follow the step-by-step setup pages. Depending on the mode, you will set the claim, pick speakers or personas, and choose the debate stage.
  4. 4 For human-led live rooms, share the debate room with the participants. For AI-led formats, launch the debate and let the system populate the turns automatically.
  5. 5 The Truth Engine scans statements as the debate unfolds. Verdicts, fallacy notes, and summaries appear alongside the live exchange.
  6. 6 Watch the live scoring and commentary areas. They update as turns are added and checked.
  7. 7 Use the challenge and follow-up tools when you want to isolate a specific claim or force extra scrutiny on a weak argument.
  8. 8 At the end, use the summary, scorecard, clip, and export tools to capture the final record.
Tip: Live Debate works great in classrooms. Students submit arguments, and the system grades them for factual accuracy in real time. No moderator needed.
What you get

Real-time verdicts on every statement. Live integrity scores per speaker that update automatically. The ability to challenge any claim mid-debate. A full post-debate report with all claims and verdicts.

📋
Feature 05
Audit Trail
See every claim you have ever verified. Filter by verdict, date, or feature. Export everything to a spreadsheet.
Who it's for Anyone tracking their fact-check history or building a record
Access level All Users
Export format CSV (opens in Excel, Google Sheets)

How to use it

  1. 1 Go to Audit Trail in the navigation bar.
  2. 2 You will see the verified-claim record for your account. The newest claims appear first.
  3. 3 Use the Verdict filter to show only True, False, Misleading, or Partially True claims.
  4. 4 Use the quick-view presets or the manual filters to narrow the record to the slice you actually want to inspect.
  5. 5 Use the Feature filter to isolate checks from claim review, debate analysis, live debate, live broadcast, or creator workflows.
  6. 6 Click any row to expand the methodology note, model breakdown, key facts, and source links for that claim.
  7. 7 Click Export CSV to download every visible row as a spreadsheet file.
What you get

A complete searchable history of every verified claim. Filters for verdict, date, and feature. Per-row model breakdowns and source links. A CSV export for analysis in any spreadsheet tool.

Part Two
Content Studio

Content Studio turns fact-checks into videos — automatically. Paste a URL, enter a topic, or run the full pipeline. The system writes the script, records the voice, finds the footage, and assembles the video.

🎥
Feature 06
Content Studio: Analyze Video
Paste a YouTube URL. The system downloads the audio, transcribes it, and fact-checks the content. Great for news clips, congressional hearings, and interviews.
Who it's for Journalists and researchers analyzing video content
Access level All Users
Supported input Any public YouTube URL

How to use it

  1. 1 Go to Content Studio in the navigation bar and choose Analyze Video.
  2. 2 Paste a YouTube URL into the input field. Copy the URL directly from your browser's address bar.
  3. 3 Click Analyze Video.
  4. 4 The system first downloads the audio from the video. Then it sends the audio to AssemblyAI for transcription. This takes 1 to 3 minutes depending on video length.
  5. 5 Once the transcript is ready, the Truth Engine extracts every factual claim.
  6. 6 Each claim gets fact-checked automatically. You can watch the verdicts fill in as each check completes.
  7. 7 Scroll through the results. Click any claim for its full model breakdown and sources.
Tip: C-SPAN videos, Senate hearing recordings, and campaign speech clips work especially well. The system handles videos up to several hours long.
What you get

A full transcript of the video. A list of every factual claim extracted from it. Per-claim verdicts with sources. An Overall Integrity Score for the video's content.

✍️
Feature 07
Content Studio: Generate Script
Enter a political topic or analysis angle. Get a full YouTube-ready commentary script — with hooks, structure, timestamps, and B-roll suggestions.
Who it's for Content creators and accountability journalists
Access level All Users
Templates Investigative · Debate Breakdown · Narrative · Educational · Conversational

How to use it

  1. 1 Go to Content Studio and choose Generate Script.
  2. 2 Enter your topic in the main text field. Be specific. For example: "Senator X's voting record on climate bills 2020-2025"
  3. 3 Pick a template from the dropdown. Each template gives the script a different structure and tone.
  4. 4 Investigative — builds tension, reveals findings one by one.
    Debate Breakdown — compares two sides with evidence.
    Narrative — tells a story from beginning to end.
    Educational — explains a topic clearly and simply.
    Conversational — direct and personal, like talking to camera.
  5. 5 Set the target video length (3 minutes, 8 minutes, 15 minutes).
  6. 6 Click Generate Script.
  7. 7 Read through the script. Each section is labeled with a suggested timestamp.
  8. 8 Edit any part of the script directly in the text area. Click Copy Script when done.
Founder Tier: If you have set up a Voice Profile, the script will match your writing style — your rhythm, your phrases, your cadence. If you have a Knowledge Base, the AI pulls research from your uploaded documents.
What you get

A full YouTube script ready to record. Sections with timestamps, hooks, and a strong close. B-roll suggestions for each section. Clean copy-paste text you can take straight to your teleprompter.

🎬
Feature 08
Content Studio: Source Video Pipeline
Paste a YouTube URL — a debate, hearing, or news clip. The system transcribes it, writes a commentary script, records a voiceover, and assembles a finished MP4 video. You do nothing except paste the URL.
Who it's for Content creators who react to and analyze existing video
Access level Founder Tier
Output MP4 video file (or MP3 audio if video assembly fails)

Pipeline stages

1. Transcript
2. Analysis
3. Script
4. Voice
5. Assemble
6. Done

How to use it

  1. 1 Go to Content Studio and choose Video Pipeline.
  2. 2 Choose the source-video pipeline path.
  3. 3 Paste the YouTube URL of the source video you want to respond to.
  4. 4 Pick a Channel from the dropdown. The channel sets which voice and style to use. Set up your channels in the Channel Manager first (see below).
  5. 5 Add Founder Context (optional) — any notes or angles you want the AI to include.
  6. 6 Click Run Pipeline.
  7. 7 Watch the stage tracker at the top. It updates every 3 seconds: Transcript → Analysis → Script → Voice → Assemble → Done.
  8. 8 When the pipeline reaches Done, a video player appears. Click play to preview your video. Click Download to save the MP4.
Channel Manager: Before you run a pipeline, set up at least one channel. Open the Channels / Voice / Knowledge page from Content Studio. Enter a channel name, pick a voice from the 11Labs voice library, and set the channel's niche (politics, news, etc.).
If video assembly fails: You still get the audio as an MP3. You can upload the MP3 to any video editor and add your own visuals.
What you get

A finished MP4 video with voiceover and B-roll. The commentary script for the video. The original transcript of the source video. All stored in your Job History for later access.

🧠
Feature 09
Content Studio: Original Topic Pipeline
Enter a topic and domain. The AI researches it using live web sources and your uploaded documents. Then it writes a script, records the voiceover, finds B-roll, and assembles a video.
Who it's for Creators who want to build original content from scratch
Access level Founder Tier
Domains Politics · Breathwork · Movement · Nutrition · Connection

Pipeline stages

1. Research
2. Script
3. Voice
4. B-Roll
5. Assemble
6. Done

How to use it

  1. 1 Go to Content Studio and choose Video Pipeline.
  2. 2 Choose the original-topic pipeline path.
  3. 3 Type your topic into the text field. For example: "How Congress votes on food safety regulations"
  4. 4 Pick a Domain from the five options: Politics, Breathwork, Movement, Nutrition, or Connection. This tells the AI which Knowledge Base and Domain Agent to use.
  5. 5 Pick a Channel for the voice and style.
  6. 6 Add optional Founder Context — angles, points, or talking points you want included.
  7. 7 Click Run Pipeline and watch the stage tracker update.
  8. 8 When done, preview the video and download the MP4.
Founder Tier: The AI uses your Knowledge Base for that domain as a primary research source — before searching the web. Your own books, papers, and notes shape the script. If you have a Domain Agent running for that domain, it may have already pre-researched recent topics automatically.
What you get

A fully assembled MP4 video on your chosen topic. An original script written from live research and your Knowledge Base. Voiceover in your channel's voice. Everything stored in Job History.

Part Three
Community & Training

Practice spotting logical fallacies. Track politicians' honesty scores over time. These features build civic literacy and public accountability.

🏋️
Feature 10
Speech & Debate Trainer
Practice speaking, argumentation, and fallacy recognition in cleaner training lanes. Use guided practice rooms, study tools, and history tracking to build judgment over time.
Who it's for Students, debaters, and anyone who wants sharper critical thinking
Access level All Users
Question bank Practice lanes, fallacy drills, and saved history

How to use it

  1. 1 Go to Speech & Debate Trainer in the navigation bar.
  2. 2 Choose the lane you want: speech coaching, debate practice, difficult conversations, fallacy study, or history review.
  3. 3 If you enter a practice room, configure the topic, structure, difficulty, or archetype on the dedicated setup page first.
  4. 4 Run the session and use the live feedback or debrief to study what held up and what broke down.
  5. 5 Use the fallacy and quiz lanes when you want shorter drills focused on recognition rather than live argument practice.
  6. 6 Review saved history afterward to track how your reasoning and delivery are improving.
  7. 7 Treat the trainer as a growth tool first. Competitive ranking is not the center of the experience right now.
What is a logical fallacy? A fallacy is a flaw in an argument. Politicians use fallacies to mislead. Learning to spot them makes you a sharper listener and thinker. The trainer explains each fallacy in plain language after every question.
What you get

A cleaner practice flow across multiple training lanes. Explanations for fallacies and rhetorical mistakes. A saved record of your sessions and debriefs. Stronger critical thinking and speaking skills with every round.

📊
Feature 11
Leaderboard (Coming Soon)
A planned public accountability tracker for politicians and public figures. The public-facing holding page is live now, but the ranking system itself is intentionally not open yet. It will launch only after the incentives and scoring logic are ready.
Who it's for The public — voters, journalists, and watchdog groups
Access level Placeholder page visible to all users
Data source Planned to use community and platform-generated claim records verified by the Truth Engine

Current status

  1. 1 Go to Leaderboard in the navigation bar.
  2. 2 You will see a clean holding page explaining that the competitive ranking surface is still being staged.
  3. 3 Use the suggested links there to move into Live Debate, Practice, or other active parts of the platform.
  4. 4 Check this manual and the product updates later for the full public ranking launch.
Note: The leaderboard is being held back on purpose. The current focus is on making debates, practice, and evidence workflows strong enough that any future ranking system rewards integrity rather than empty performance.
What you get

A clear explanation of the feature's current status. Direct links into the live parts of the platform. A cleaner future launch path for public ranking once it is ready.

Part Four — Founder Tier Only
Founder Tools

These features are available on the Founder tier only. They let you teach the AI your voice, build a research library, and run automated agents that create videos on a schedule.

🎤
Feature 12
Voice Profile
Teach the AI your writing style. Upload samples of your writing and notes about how you like to sound. Every script the AI generates will match your rhythm and cadence.
Who it's for Creators who want AI-generated scripts that sound like them
Access level Founder Tier
Input types Text samples, style notes, phrases to avoid

How to use it

  1. 1 Go to your Founder Dashboard. Click Voice Profile.
  2. 2 Paste 3 to 10 samples of your writing into the writing samples field. Use scripts, blog posts, essays, or social media posts you have written before.
  3. 3 Fill in the Style Notes field. Describe how you write. For example: "I use short sentences. I ask rhetorical questions. I build to a strong close."
  4. 4 Add Phrases to Avoid — words or expressions you never use. The AI will not include them in your scripts.
  5. 5 Click Save Voice Profile.
  6. 6 From now on, every script the Generate Script and Video Pipeline tools create will match the style you described.
Tip: The more writing samples you add, the better the match. Update your Voice Profile whenever your style evolves. You can save multiple profiles and pick one per channel.
What you get

AI-generated scripts that sound like you wrote them. Consistent voice across all your channels. The ability to swap between voice profiles for different content styles.

📚
Feature 13
Knowledge Base
Upload books, papers, research, and your own writing — per domain. When the AI generates a script, it reads your documents first. Your knowledge shapes every video.
Who it's for Subject-matter experts who want the AI to use their research
Access level Founder Tier
Supported file types PDF, DOCX, PPTX, TXT

How to use it

  1. 1 Go to your Founder Dashboard. Click Knowledge Base.
  2. 2 Pick a domain: Politics, Breathwork, Movement, Nutrition, or Connection.
  3. 3 Click Upload Document. Select a PDF, DOCX, PPTX, or TXT file from your computer. The system accepts files up to 50 MB.
  4. 4 Add a short description so you can find the document later. For example: "Senate voting records 2024, downloaded from congress.gov"
  5. 5 Click Save. The system indexes the document. This takes under a minute.
  6. 6 Repeat for as many documents as you want. There is no hard limit per domain.
  7. 7 Now run the original-topic pipeline or Generate Script for that domain. The AI searches your documents before it searches the web.
What to upload: Books and research reports give the AI depth. Your own writing gives it your perspective. Government data files and news archives give it current facts. Upload a mix of all three for the best results.
What you get

Scripts that draw on your research, not just public web sources. Consistent, authoritative content rooted in documents you trust. A searchable library of your knowledge, organized by domain.

🤖
Feature 14
Domain Agents
Autonomous AI agents that run on a schedule. Each agent researches current topics in its domain, then generates and assembles finished videos automatically. You wake up to new content — without lifting a finger.
Who it's for Founders who want automated, scheduled content creation
Access level Founder Tier
Agents One per domain: Politics · Breathwork · Movement · Nutrition · Connection

How to use it

  1. 1 Go to your Founder Dashboard. Click Domain Agents.
  2. 2 Pick a domain from the list. Each domain has its own agent card.
  3. 3 Click Configure Agent on the card you want to activate.
  4. 4 Set the schedule: daily, every 3 days, or weekly. Pick the time of day for the agent to run.
  5. 5 Pick a channel for the agent to publish to. The agent uses that channel's voice and style.
  6. 6 Add Topic Guidance — keywords or themes you want the agent to stay focused on. For example: "federal budget, congressional voting, campaign finance"
  7. 7 Toggle the agent to Active. Click Save.
  8. 8 The agent runs on its schedule. Each run creates a new pipeline job. You will see it in your Job History when it finishes.
  9. 9 Review the finished video in Job History. Approve it before publishing, or set the agent to auto-publish if you prefer.
Tip: Pair each Domain Agent with a Knowledge Base for that domain. The agent will include your research in every video it creates. Set up your Voice Profile too — all agent-generated scripts will match your style.
Full automation stack: Voice Profile + Knowledge Base + Domain Agent = videos that sound like you, cite your sources, and publish on a schedule — with no manual work after setup.
What you get

Finished videos created automatically on a recurring schedule. Original research drawn from live web sources and your Knowledge Base. Scripts in your writing voice. A growing library of content — built while you sleep.

Quick Start
Getting Started in 5 Minutes

New to RhetoricalPoints? Follow these steps to run your first fact-check right now.

First fact-check

  1. 1 Go to rhetoricalpoints.com and click Sign Up.
  2. 2 Create your account with an email and password. No credit card needed for the free tier.
  3. 3 You land on the Dashboard. Click Truth Engine in the navigation bar.
  4. 4 Copy a claim from any news article or politician's social media post.
  5. 5 Paste the claim and click Verify Claim.
  6. 6 In about 20 seconds you will see your first verdict with sources. That is it.
Free tier includes Truth Engine, Speech / Transcript Checker, Debate Analyzer, Live Debate, Audit Trail, Speech & Debate Trainer, and the current Leaderboard placeholder page
Founder tier adds Content Studio Pipelines, Voice Profile, Knowledge Base, Domain Agents