Institutional prediction market infrastructure.
Software, APIs, and data for funds, corporates, and trading desks operating on Kalshi and Polymarket.
Cross-venue research, normalized intents, risk-gate primitives, append-only portfolio ledger, and reconciliation-ready read models. Prime-brokerage-style software for prediction-market sleeves — across Kalshi and Polymarket, wired through BYOK credentials the desk owns.
Edward Lloyd's coffee house, 1688 — the original institutional infrastructure for systematic risk.
The concrete call
A representative buyer-side call, in three beats: the exposure, why the in-house desk does not absorb it, and the delegation. The page below is the institutional surface built to take exactly this call.
The exposure
A commodities PM is carrying $100M of nickel exposure marked off LME. The unhedged sleeve is not nickel beta — it is event-risk basis the LME cannot isolate: USTR Section 232 action on processed Indonesian nickel, an Indonesian export-quota revision, a Philippine mining moratorium. The same book's adjacent copper line sits on a Peru runoff and Chilean policy reform. LME and SHFE hedge the metal. Kalshi and Polymarket are where the events price.
Why the desk does not absorb it
That hedge does not belong on the LME desk. The metals trader is not going to KYC into a US-regulated event exchange. The risk officer is not going to let venue credentials sit on a trader workstation. The fund administrator is not going to ingest a one-off CSV format into the NAV pipeline. The PM has the view; the desk does not have the operational surface to express it.
The delegation
So the PM delegates the prediction-market leg — operationally closer to FCM clearing, give-up execution, or a securities-lending borrow than to an FX algo. The thesis, the size, and the venue relationship stay with the desk. The software is what gets handed off: normalized data feed, intent shape, risk gates, portfolio ledger, and reconciliation-ready read models. That delegation is what this layer is. SimpleFunctions is the OMS/EMS plus portfolio control plane; capital never leaves the desk's own Kalshi or Polymarket account.
The institutional pipeline — six steps
Every desk that runs on this institutional prediction market infrastructure flows through the same six-step pipeline. Research feeds views; views shape intents; intents pass risk gates; gates clear into venue routes; recorded fills land in the portfolio ledger.
01Research
Pull world-state snapshots, screen ~9,000 normalized markets, query gov + econ datasets — all through one institutional prediction market infrastructure surface
02View
Encode the desk thesis as a directional + sized view; views feed downstream signals and seed risk-gate parameters
03Intent
Declare what the desk wants done: market, side, size, price, trigger condition, expiry — idempotent and replay-safe across both venues
04Risk gate
Pre-trade checks against operator-configured limits — size, exposure, drawdown, regime, daily-loss, dry-run toggle
05Venue route
Map normalized intent to Kalshi or Polymarket order shape; submit through the venue connection bound to the desk BYOK credential
06Reconcile
Read recorded fills, positions, risk, and attribution from the portfolio ledger; venue settlement import is connected as an integration step when required
Who uses institutional prediction market infrastructure
Five operating profiles, three surface depths. Each desk picks the surface that matches its mandate, ops capacity, and venue relationship. Capital stays at the venue across every profile.
Ledger and audit
Four product facts a desk operator integrates against. The portfolio ledger is the contract between the platform and the desk's accounting pipeline.
Portfolio ledger
Intents, risk-gate decisions, venue submissions, recorded fills, and reconciliation events are written append-only when they pass through the platform — with actor, timestamp, source, and attribution confidence.
Read scopes
Data surfaces (markets, candles, indicators, world state) require no venue credential. Execution surfaces act only when a BYOK key is bound to an authenticated session; every action is logged against that session.
Authenticated reads
/api/portfolio/fills, /api/portfolio/attribution/grouped, /api/portfolio/positions, and /api/portfolio/risk expose the ledger to the desk.
Export
CSV / JSON / Parquet export wired into the desk's accounting pipeline as an integration step. Venue statements remain the source of truth for capital movement.
Data and execution surfaces
Eight production surfaces, all part of the same institutional prediction market infrastructure. Pick the surface that matches the workflow; every surface shares the same portfolio ledger and BYOK posture.
Onboarding path
Three phases, operator-paced. Sandbox is immediate, dry-run is one business week, live is paced by the desk's own soak time. No phase requires SimpleFunctions to hold funds or take a venue credential the desk does not own.
01Sandbox
Read-only access to data, world state, indicators, gov + econ overlays. Test integrations against the public API surface — no venue credential, no capital, no commitment.
02Dry-run
Bind a Kalshi or Polymarket key in dry-run mode. Every intent flows the full pipeline — risk gates, normalization, route — but stops short of submission. Validates the operator's policy without capital at risk.
03Live
Flip dry-run off on a per-strategy basis. Risk gates, portfolio ledger writes, and configured reconciliation reads come live; fills reach the venue and settle to the desk's existing account.
Read next from the library
Matched from SimpleFunctions blog, opinions, technical guides, concepts, and learn pages.
From Boesky to Bots: Porting Hedge Fund Alpha to Prediction Markets
Classic hedge fund strategies map onto Polymarket and Kalshi with surprising fidelity. Ten mappings from cleanest analog to most speculative — Levy, Buffett, Greenblatt, Tartaglia, Griffin, Meriwether, Soros, Taleb — with documented cases, dollar amounts, and the 2024–2026 academic evidence.
Why "Prediction Market Index Funds" Are Mathematically Dubious
Index funds need continuous returns, shared factor exposures, and meaningful weights. Binary prediction-market contracts have none. A naive PM index converges to noise, not a return.
Kalshi vs Polymarket: Which Prediction Market Should You Trade?
In-depth comparison of Kalshi and Polymarket for prediction market traders. Regulatory structure, liquidity, fees, API tooling, and cross-venue trading with SimpleFunctions.
Piping prediction market signals into your existing trading system
Three integration patterns for piping Kalshi and Polymarket data into existing trading infrastructure: cron polling, agent middleware, and thesis-as-filter.
Event Contract
Event contracts are binary instruments tied to real-world events. Learn how they work on Kalshi and Polymarket.
Cross-Venue Edge Detection: Kalshi vs Polymarket
How to detect and exploit price divergences between Kalshi and Polymarket. Covers why pure arbitrage fails, cross-venue edge detection algorithms, and thesis-informed trading strategies.
FAQ
What is institutional prediction market infrastructure?
The software, APIs, and data layer between a fund or trading desk and the underlying prediction market venues (Kalshi and Polymarket). It covers research surfaces (probability, screening, indicators), normalized intents and triggers, risk-gate primitives, portfolio ledger reads, and reconciliation-ready data models — without taking custody of the desk's capital. SimpleFunctions is the institutional prediction market infrastructure layer for Kalshi and Polymarket.
Who actually uses this?
Hedge funds running macro or event-driven mandates, family offices holding cheap binary tail exposure, bank rates and FX desks running cross-asset overlays, corporate treasuries hedging policy uncertainty, and prop desks running systematic event-contract strategies. Each surface (data only, full pipeline, hosted autonomous) maps to a different operating posture.
How does this differ from going direct to Kalshi or Polymarket?
The venue gives you a market, an order endpoint, and an account. It does not give you cross-venue normalization, an idempotent intent shape, regime-aware risk gates, ledger-backed portfolio reads, reconciliation-ready data models, or a real-time data surface designed for agents. Institutional prediction market infrastructure is the missing software between raw venue API and fund operations.
How does this compare to prime brokerage?
Prime-brokerage-style software for prediction-market sleeves — data, intents, risk gates, audit, reconciliation — without the prime-brokerage relationship. Customers route through their own Kalshi and Polymarket accounts via BYOK.
How do fees work?
Tiered software subscription per seat or per agent, plus optional usage-based pricing for high-volume API and execution workloads. Pricing is set after a technical review — fund mandate, expected request volume, and desk-specific surfaces (audit retention window, BYOK key count, dedicated infrastructure) all factor in. Capital remains at the venue; SimpleFunctions never charges trading commissions.
What does the audit posture look like?
Every intent, risk-gate evaluation, venue submission, recorded fill, and reconciliation event that passes through SimpleFunctions is written append-only with actor, time, source, reason, and attribution confidence where available. Export format and retention are desk-integration details; venue statements remain the source of truth for capital movement.
How long does institutional onboarding take?
Sandbox is immediate — read-only data access requires no credential exchange. Dry-run typically lands within one business week, long enough to bind a venue key, configure risk gates, and run the desk's intended policies end-to-end without capital. Live cutover is operator-paced: most desks soak in dry-run for one to two weeks before flipping individual strategies live.
Where does the capital sit?
In the desk's existing Kalshi or Polymarket account, accessed via BYOK credentials the desk owns. The platform consumes the key within an authenticated session for the operations the desk has approved.
Is there a sandbox?
Yes. The data and research surfaces (markets, candles, indicators, world state, gov + econ overlays) are reachable through the public API with rate-limited free access — no venue credential required. Execution surfaces ship a dry-run mode for full-pipeline rehearsal before any real fill is attempted.
How does reconciliation work for fund accounting?
The current first-party API exposes recorded fills, position snapshots, daily attribution, grouped attribution, activity, and risk snapshots. Venue settlement import and external accounting export are integration steps layered on top of that ledger, not assumed for every account by default.
What about NAV reporting and fund administration?
For desks that wire an export, the portfolio ledger/read models can feed standard accounting or compliance pipelines. NAV calculation itself remains with the fund administrator; venue statements and venue-side timestamps stay authoritative for capital movement. No proprietary NAV format is imposed.
Can the desk bring its own model and risk policy?
Yes. BYOM (bring-your-own-model) and BYOR (bring-your-own-risk) are first-class. The intent shape and risk-gate primitives are public; desks can author models and risk rules in their language of choice (Python, Go, TypeScript, Julia) and pipe them through the same execution surface. The hosted Portfolio Autopilot is one of several runtimes, not a requirement.
Related surfaces
Hedge fund prediction market
Buyer-specific framing — alpha discovery and tail hedging on event contracts.
Prediction market funds
For funds raising capital to deploy systematically into Kalshi and Polymarket.
Prediction market execution
Intents, triggers, routing, monitoring — Kalshi and Polymarket execution workflows.
Prediction market hedging
Map real-portfolio exposures to event contracts as binary or basket hedges.
Real-time data API
Sub-second WebSocket and REST surface — the live feed that triggers actually watch.
Agentic CLI
Operator and agent control plane for scripted research, intents, reconciliation.
Portfolio Autopilot
Hosted LLM-driven autonomous policy on top of the same intent + risk-gate stack.