Prediction Market Analytics

See what’s actually
likely to happen.

ChatterLab simulates how events play out — with thousands of AI agents modelling real dynamics — to produce probability estimates independent of market consensus.

1
Graph
2
Agents
3
Simulate
4
Report
5
Done
Scroll to start simulation
 
Built for anyone who wants to know what’s actually going to happen
Polymarket & Kalshi compatible Base rate analysis Multi-agent simulation Independent estimates

Three steps to a better estimate

Paste a question. Get an analysis. It’s that simple.
01 — DESCRIBE
Drop a link or describe the event
Paste a Polymarket or Kalshi link, or just describe the question in your own words. The copilot figures out the resolution criteria, timeline, and key drivers.
02 — SIMULATE
Thousands of agents model how it plays out
Policymakers, analysts, journalists, economists — each with their own expertise and biases — react to the scenario and each other. Cascade effects emerge naturally.
03 — ANALYSE
See what the data says
Get a probability estimate with confidence range, key drivers ranked by impact, base rate analysis, and a clear view of where the model diverges from market consensus.

Any event with a yes-or-no outcome

If there’s a prediction market for it, ChatterLab can simulate it.
01

Politics & Geopolitics

  • Election outcomes
  • Policy decisions
  • International agreements
  • Leadership changes
02

Monetary Policy & Macro

  • Rate decisions
  • Inflation targets
  • Employment thresholds
  • Central bank actions
03

Regulatory & Legal

  • Court rulings
  • Regulatory approvals
  • Legislation passage
  • Enforcement actions
04

Tech & Crypto

  • Product launches
  • Protocol upgrades
  • Adoption milestones
  • Market structure changes
05

Science & Climate

  • Research milestones
  • Climate targets
  • Health outcomes
  • Space events
06

Custom Questions

  • Any yes/no question
  • Corporate events
  • Cultural milestones
  • Niche markets

Talk to the copilot

Ask about any prediction market question. The copilot will walk you through setting up the analysis.

How it thinks

The copilot doesn’t just pattern-match — it builds a simulation with real dynamics. Here’s what makes it different:

  • Starts from base rates — how often has this type of thing actually happened?
  • Maps the key drivers — who and what actually moves the needle
  • Simulates cascade effects — how one event triggers another
  • Produces an estimate independent of market consensus
  • Shows where the model diverges from market pricing and why
Start analysing
ChatterLab Markets
Paste a Polymarket or Kalshi link, or just describe the event you want to analyse. I'll set up a simulation that estimates what's actually likely to happen.

Common questions

How is this different from just asking ChatGPT?
ChatGPT gives you one model's opinion. ChatterLab simulates thousands of agents with different expertise, biases, and information — modelling how events actually play out through a complex system. The probability comes from the dynamics, not from pattern matching.
Where does the probability estimate come from?
From the simulation. Thousands of AI agents — modelled on real actor types like policymakers, analysts, and journalists — play out the scenario over multiple rounds. The model-implied probability reflects how the scenario resolves across all simulation paths.
Is this financial advice?
No. ChatterLab is a pure analytics platform. All outputs are observations — what the model sees, not what you should do. We describe probabilities and divergences, never recommendations. For informational purposes only.
What markets can I analyse?
Any event with a clear yes-or-no outcome. Works with Polymarket, Kalshi, or any custom question. Politics, monetary policy, regulatory decisions, tech, crypto, sports — if you can define what YES means, ChatterLab can simulate it.
How accurate are the estimates?
We're honest about uncertainty — every estimate comes with a confidence range. The model is calibrated against historical base rates, meaning when it says 60%, it's right about 60% of the time. But no model is perfect, and we say so.
Can I connect my own data sources?
Yes. You can plug in RSS feeds, API endpoints, or any data source. The copilot can also pull live market pricing data. All external content is automatically sanitised to prevent manipulation.

Ready to see what the data says?

Sign up for early access. We’re onboarding analysts now.

No spam. We notify you when broader access opens and nothing else.

Stop guessing. Start simulating.

Get probability estimates backed by multi-agent simulation — not vibes.