How it works
How LLM Peers works—from question to decision brief
This page is the practical guide. Every run follows a structured deliberation path. You choose Free or Premium Research, optionally pick a research panel and attach My sources on Premium, then watch independent models answer, peer-rank, and synthesize one recommendation.
1. Sign in and ask one difficult question
Sign in with Google. Bring a real decision: a choice with trade-offs, constraints, and consequences. Vague curiosity still works, but the product shines when “good enough chat” is not good enough for the commitment you are about to make.
2. Choose Free or Premium Research
There is no Free / Pro / Max subscription. Free Research gives you welcome runs when you join, then one Free Research every 24 hours with an efficient multi-lab council. Premium Research uses prepaid wallet credits for research panels, custom models, deeper depth, My sources, and frontier labs—charged as model usage plus a platform fee.
Free Research
Daily multi-model deliberation with efficient models. Same stages as Premium—clarify, research, review, decide—without spending credits when Free Research is available.
Premium Research
Start from a curated research panel (Balanced default by default) or customize peers and chairman. Attach files and links from My sources. Choose Standard, Deep, or Maximum depth. Pay from your wallet only for what you run.
3. On Premium: pick a research panel (or customize)
Premium Research defaults to a research panel matched to your decision—industry boards (healthcare, manufacturing, retail, finance, and more), role panels (career, hiring, procurement), or decision modes (adversarial, budget, maximum rigor). Accept the panel, or customize peers, chairman, and depth before confirming. Free Research skips this step and uses a fixed efficient lineup.
4. The deliberation path
- 01
Clarify with intake when needed
An optional intake step may ask focused clarifying questions (single choice, multi choice, or short text). Answers enrich the research brief. You can skip if context is already complete.
- 02
Gather web context
When web research is enabled, the board collects a current briefing from public sources so models can reason with fresher external context alongside your question.
- 03
Attach My sources (Premium)
Upload PDFs, Word, PowerPoint, Excel, text, images, or paste public links in My sources. Attach Ready items so every peer and the chairman share a private, citable briefing.
- 04
Run the AI council
Multiple models produce independent answers—recommendations, risks, assumptions, and next steps—without seeing each other’s work. This is Stage 1 of deliberation.
- 05
Peer review and rank
Answers are anonymized and ranked against shared evaluation criteria. Stage 2 surfaces comparative strength, agreement, and dissent without brand-driven preference.
- 06
Chairman synthesis
A lead agent synthesizes rankings and evidence into one decision brief: recommendation, ranked options, assumptions, risks, and next action. Stage 3 is the artifact you take forward.
5. Inspect progress and the brief
Progress is written stage by stage to your research conversation so the UI updates as each phase completes. You can inspect intake, web context, My sources citations, council answers, peer rankings, and the final synthesis rather than trusting a black-box reply. On Premium runs, a billing receipt shows model usage and the platform fee.
After the run
- Read the decision brief and supporting stages in your workspace
- Open [Doc: …] citations to jump to matching My sources passages
- Export or share a public research link when others need the same case
- Return to conversation history for related follow-up decisions
- On Premium, review the receipt and wallet balance before the next run
- Challenge assumptions that would change the recommendation before you commit
Tips for stronger runs
- State the decision, not only the topic (“Should we X or Y under constraint Z?”)
- Include constraints early: budget, timeline, risk tolerance, non-negotiables
- Name what “good” looks like so ranking criteria can stay grounded
- Use Free Research for everyday consequential questions; switch to Premium for high-stakes or specialized panels
- Start Premium from Balanced default unless a specialized panel clearly fits
- Enable web research for time-sensitive market or vendor questions
- Attach My sources when the answer depends on your decks, sheets, or private docs
How long does a run take?
Longer than a single chat reply, because multiple models and stages execute. Exact timing depends on question complexity, web research, model latency, and Premium depth. You can watch progress as stages complete—and leave; you can be emailed when research is ready.
Do I need integrations to start?
No. Sign in with Google, ask a question, and run Free or Premium Research. No CRM, wiki, or data-warehouse integration is required to begin.
What if Free Research is used for today?
Wait for the next Free Research window, or switch to Premium Research with wallet credits to continue immediately with panels and advanced models.
Can I interrupt or refine mid-flight?
Intake is the structured clarification step before the council. After completion, start a follow-up research question with tighter constraints or new evidence.
Dive deeper
- Free & Premium ResearchBilling modes and wallet credits
- Research panelsCurated Premium starting lineups
- My sourcesFiles, images, and links for Premium grounding
- ModelsLabs and model catalogs
- AI CouncilIndependent multi-model answers
- Peer reviewAnonymous ranking across models
- Decision briefChairman synthesis and next actions
- PricingFree Research and example Premium charges
