Forecast AI
Forecast AI runs seven specialized agents to produce calibrated probability forecasts and structured trade recommendations.

Forecast AI: Open Source AI Forecasts For Polymarket And Kalshi
What is Forecast AI?
Forecast AI runs seven specialized agents in parallel to produce calibrated probability forecasts for prediction market events on Kalshi and Polymarket, with trade recommendations formatted for Robinhood's Agentic Trading MCP.
Ready to try Forecast AI?
Open the official site and get started in a few clicks.
Forecast AI Overview
Forecast AI is an open source multi-agent system that converts raw market signals into calibrated probability forecasts for prediction markets. It runs seven specialized agents in parallel, each focused on a distinct data stream. A Consensus Engine takes their independent outputs and produces a single probability, confidence score, and structured trade recommendation.
Single-model LLM prompts suffer from recency bias and poor handling of conflicting evidence. Forecast AI fixes this by separating concerns. Each agent works in isolation on its domain, then the Consensus Engine applies historical accuracy weights and Bayesian averaging. The agents cover news feeds, social sentiment, Reddit discussions, structured research, macro indicators, on-chain activity, and live order books from Kalshi and Polymarket. Every output includes the reasoning trail from each agent, so you can audit how the final number was reached.
Active traders who want an evidence-ranked edge will find it useful. Developers can extend or embed it as a ready-made intelligence layer. Operators who prefer to keep API keys and compute under their own control can self-host the full stack. Forecast AI is a research layer that surfaces probabilities, and the user handles all execution.
What sets it apart is measured performance. Public claims state a +24 percent improvement in Brier score over a vanilla GPT-4 baseline. The Robinhood Recommendation Engine converts consensus into a structured prompt with the action, target contract, and risk notes for the Robinhood Agentic Trading MCP endpoint. Modular LLM providers support OpenAI with automatic fallback to Gemini, Anthropic, OpenRouter, or local Ollama. The core code is fully open under Apache 2.0, and a live dashboard at forai.tech shows real-time consensus against market midpoints.
Forecast AI Key features
-
Seven Specialized Agents
Seven independent agents analyze distinct data streams like news and market pricing. Each agent works in isolation on its specific domain to avoid recency bias.
-
Consensus Probability Engine
The engine aggregates probabilities from all agents and applies historical accuracy weights. It outputs a single calibrated probability with uncertainty bounds and a short rationale from each agent.
-
Historical Vector Memory
The system stores prior forecasts and outcomes as vector data. It uses this history to improve weightings over time and maintain context across markets that run for a long time.
-
Robinhood Trade Prompts
This engine converts the consensus probability into a structured trade prompt. Users can paste this prompt directly into a client connected to the Robinhood Agentic Trading MCP.
-
Public Read API
The API serves live consensus data for topics like crypto and macro. A monthly compute budget guard and rate limits protect the public feed.
-
Modular LLM Providers
Users can choose from multiple LLM providers like OpenAI and Gemini. If one provider fails, the system automatically retries the next one.
-
Terminal And Server Access
Developers can run one-shot predictions or continuous watch loops from the terminal. They can also operate the system as a microservice through the FastAPI server.
Forecast AI Fees
-
Platform fee The core open source software carries no platform fee.
-
LLM API costs Users supply their own LLM API keys and pay the respective providers directly.
-
FactsAI research calls Optional FactsAI deep research calls cost 0.012 USD per request.
-
Hosted dashboard tiers Free accounts receive limited analyses with standard models, while $FORAI holders unlock higher limits, deeper runs, and priority features.
-
Trading fees Forecast AI charges no trading fees because execution happens inside the user's own Robinhood session.
How to use Forecast AI
-
Clone the repository
Copy the codebase from the official GitHub repository.
-
Run the setup command
Execute the interactive setup script to configure your LLM API keys and environment variables. You can also add an optional FactsAI key for source-backed research.
-
Start the monitoring loop
Launch the public feed or a continuous watch loop using the command line interface. This begins background tracking.
-
Generate a forecast
Provide a market query or identifier to generate a one-shot prediction.
-
Review the consensus report
Read the final probability, confidence score, and agent reasoning. The system filters anomalies and applies calibration to the data.
-
Generate the trade recommendation
Convert the consensus output into a structured trade prompt for Robinhood.
-
Execute or discard the trade
Paste the prompt into an authenticated client connected to the Robinhood Agentic Trading MCP. You must manually approve or reject the action inside your own account.
Forecast AI Screenshots
Is Forecast AI safe & legit?
<strong>Forecast AI</strong> is an open source project built by a developer known as <strong>codebyollie</strong>, with the full codebase public on GitHub under the Apache 2.0 license. No specific red flags or security incidents have been reported, and the tool never takes custody of your funds or places trades on your behalf. The project does have an associated <strong>$FORAI token</strong>, so you should always verify the official contract address on the project website before any token interaction, and keep in mind the system is still in active development.
X account intel @forecast_agents
-
Based in Europe
-
Joined X October 2025 9 months ago
-
Username changes 7 renames last on Jun 30, 2026
-
Signup source Eastern Europe (Non-EU) App Store
Public profile signals from X. Use as one input alongside other due-diligence.
Wallet blacklist scan checked Jul 25, 2026
- MetaMask Not flagged
- Phantom Not flagged
- ScamSniffer Not flagged
- EtherAddressLookup Not flagged
- Keplr Wallet Not flagged
Domain forai.tech checked against public crypto wallet blacklists.
Forecast AI Updates
-
Happy Saturday again. What we're working on now and what's coming next::
Working on now: polishing the profile/user experience - wallet linking, real badges, holder tiers, and the foundation for everything tied to your account.
Building next: the big one.
Run full Forecast AI analysis directly from our official website (http://forai.tech) . No install, no local setup - choose a market and let the agents work.
Access will scale with $FORAI holdings:
Free: a limited number of analyses, standard models, profile + basic history.
Holders: more analyses, deeper agent runs, access to more powerful models, custom market search & analysis presets, and priority access to new agents and data sources.
Then we start building the layer around it: cloud forecast history, public track records, reputation & badges, and eventually forecast competitions.
The idea is simple - your profile shouldn’t just show what you hold. Over time, it should show what you predicted and how good you are at it.
The core Forecast AI infrastructure stays open source. Holder utility is being built on top of it, not instead of it.
Building piece by piece. Will keep sharing as each part gets closer.
Github: https://github.com/codebyollie/forecast-agents
Stay safe fam and always do research.
- 5 replies
- 2 reposts
- 19 likes
- 567 views
-
- 6 replies
- 23 likes
- 3.6K views
-
Forecast AI × Polyfactual 🤝
We’ve successfully integrated @factsdottrade Deep Research API by @polyfactual directly into Forecast AI.
The integration brings:
→ Deeper and high-quality research directly into our Research & Macro agents
→ Source-backed answers with citations
→ Full integration guide for developers and self-hosters in our open-source GitHub repo
Deep research → agents → consensus → forecast.
Now live and integrated into the Forecast AI stack.
More integrations coming.
- 4 replies
- 3 reposts
- 32 likes
- 8.2K views
-
HM! Finally - the Forecast AI Telegram community is open.
https://t.me/+_QfHgas_OrplNDBi
Welcome everyone. This will be the place to talk directly with us, share feedback and ideas, discuss prediction markets, AI agents, $FORAI, and follow what we’re building in real time.
https://t.me/forecast_agents (our Telegram channel link)
Still early. A lot to build. More updates coming today. One of them is almost ready - been working on an integration that adds another layer to what our agents can do. Not revealing this one just yet- 1 replies
- 8 likes
- 3.1K views
-
Official Forecast AI Telegram group is here.
http://t.me/forecast_agents
Start joining - this will be the main place for all updates together with @forecast_agents X account.
I’m setting up the channel now and need to configure everything before fully opening it up.
Going live today. Stay tuned- 8 likes
- 2.2K views
-
- 2 replies
- 6 likes
- 1.1K views
-
Quick update.
Hoping to ship one more Forecast AI update today. Can’t share much about this one yet…
Taking a quick break, then back to testing and building.
Also, the official TG group goes live today. Time to bring the community together and build alongside you.
Stay tuned.- 1 reposts
- 5 likes
- 790 views
-
New Forecast AI update is live.
User profiles & identity are now available on http://forai.tech.
(actively testing this now)
You can now sign in with email or a Web3 wallet powered by @privy_io and create your Forecast profile.
What’s live:
→ $FORAI on-chain holder verification with Free / Holder / Pro tiers
→ Account badges, starting with Early Adopter and holder credentials
→ Infrastructure for building your forecasting track record & reputation
We also pushed a maintenance + documentation update to GitHub: http://github.com/codebyollie/forecast-agent
→ Fixed several backend edge cases for more reliable agent execution
→ Expanded the README with CLI references, architecture and deployment docs
→ Simplified self-hosting for Render, Railway and local environments
More shipping soon.
- 1 replies
- 5 likes
- 1.6K views
-
Just pushed a new Forecast AI update. 🛠️
The backend for user profiles + $FORAI holder tiers is now live.
https://github.com/codebyollie/forecast-agents
What’s ready:
→ User authentication powered by Privy
→ Persistent profiles and linked wallets
→ On-chain $FORAI balance verification on @RobinhoodApp chain.
→ Automatic Free / Holder / Pro Holder tiers
→ Holder, Pro Holder & Early Adopter badges
This is the foundation for the utility layer shared earlier. different levels of access, compute, agent capabilities, Next: bringing profiles to http://forai.tech so you can actually connect, see your tier and be ready for what's coming next.
Backend shipped. Frontend now in progress.
Will keep you updated.
- 8 likes
- 744 views
-
Been thinking a lot about how $FORAI utility should actually fit into Forecast AI.
One thing won’t change: the core stays open source.
Anyone can download Forecast AI, bring their own API keys, and run the agents themselves.
But not everyone wants to install, configure, and maintain AI infrastructure.
So the next thing I’m building is the user layer directly on http://forai.tech.
Every user will be able to create a profile and run Forecast AI directly from our website.
Free Tier - no $FORAI required
→ Create your Forecast profile
→ Run a limited number of analyses
→ Access the standard 7-agent setup
→ Build your forecast history
→ Track accuracy, reputation, and earn badges
$FORAI Holder Tier
→ More analyses and higher compute limits
→ Access to our most capable LLMs
→ Deep agent runs with more sources and additional consensus rounds
→ Search and analyze the markets you actually care about
→ Build your own analysis presets
→ Cloud-saved presets and forecast memory
→ Early access to new agents, models, data sources, and features
Further down the roadmap, I also want to explore public forecasting profiles, leaderboards, and weekly competitions where the best forecasters can earn $FORAI or any other assets.
This is the direction I’m exploring - not the final utility spec yet. Would love to hear feedback before we lock the roadmap.
- 4 replies
- 6 likes
- 621 views
Forecast AI Reviews & Comments
No reviews yet. Be the first to share your experience.
Forecast AI alternatives
Other Polymarket tools that do a similar job. Compare and pick what fits your workflow.
-
Compare
Poly Prediction
Traders who want context behind Polymarket prices get AI-driven analysis that connects li…
-
Compare
Ember
Traders who want a second opinion on Polymarket odds get daily AI probability assessments…
-
Compare
Polyfactual
Polyfactual combines live prediction market data with AI to give traders and developers i…
-
Compare
Insiders bot
Insiders.bot tracks profitable Polymarket wallets and lets traders copy their trades auto…
Recommended reading
In-depth guides for AI Tools — get more out of tools like Forecast AI.
-
Beginners
Polymarket Weather Markets: A Data-First Trading Guide
Learn how to trade Polymarket weather markets with a repeatable workflow: read the station rules, stack ECMWF and METAR forecasts…
8 min read -
Market News
Top 20 Polymarket Influencers And Traders to Follow on X (2026)
A ranked guide to 20 Polymarket influencers on X. Find data analysts, political sharps, and niche traders who share their reasoni…
6 min read -
Tool Reviews
Free Polymarket Open Source Tools and Trading Bots on GitHub
Free Polymarket open source tools and trading bots on GitHub. Datasets, backtesters, copy trading, weather bots, and AI agents, w…
10 min read
Ready to try Forecast AI?
Open the official site and get started in a few clicks.





Leave a comment