NyquistNyquist
PILOT PROGRAM · BY APPLICATIONStart a 60-day pilot →
★ The agent research desk for funds under $500M AUM   /   42 domains · 54 hubs

Give it a ticker.Get back aweek of work in minutes.

For the PM who is also the risk desk. A swarm of agents reads the filings, argues the bull and the bear case, and hands you a call with the reasoning shown. Open any step and check the work yourself — it's not a chatbot or a black box.

★ No signup, no call · the demo opens empty, you enter your own book in your browser
★ Reasoning shown · the why behind the call, not just the verdict
★ Runs overnight · open each morning to what moved on your watchlist
★ Built by quants & engineers · founder-led, hands on the code · reads on the method ↗
★ Pilot program · by application · design partners wanted
★ 01   Platform architecture

Vertical integration down to the schema, not an AI wrapper.

Generic AI tools bolt a chat window on top of third-party spreadsheets and market feeds. Nyquist builds the complete trading and risk stack from the mathematical substrate up: a typed bitemporal ontology, a high-performance quant engine, and an autonomous agent swarm operating as native tools over verified financial data.

Layer 01

Typed Bitemporal Ontology

54 core financial entity types with strict relational typing. Dual time-axes separate Valid Time (when a market price, curve, or trade was true in the world) from System Time (when it was recorded), so any figure can be recomputed as of a past date and traced to the inputs and run that produced it.

Data Ground Truth →
Layer 02

Complex Quant Compute Engine

A deterministic mathematical core spanning 34 derivative model families (stochastic volatility, jump diffusions, finite difference PDEs, American option trees), yield curve bootstrapping, MBS prepayment, ABS cashflow waterfalls, Extreme Value Theory (EVT), and bilateral XVA netting.

Deterministic Math →
Layer 03

Autonomous Agent Research Desk

36 specialized analyst agents on named investing methodologies (moat-quality, macro-asymmetry, tail-risk, factor allocation) debating theses in structured committees. Agents execute tools directly against the quant core — producing auditable calls with explicit invalidation triggers.

Tool-Grounded Swarm →
★ 02   Domain coverage

One engine across the entire balance sheet.

Nyquist spans 42 financial domains and 54 analytical hubs on a single codebase — connecting pricing, risk, balance sheet management, and research without brittle exports between vendors.

★ 01 · Pricing & Derivatives

Multi-asset valuation

From liquid rates and equities to complex exotics and structured credit. A single deterministic engine computes closed-form solutions, PDE grids, and volatility surfaces without external black boxes.

★ 34 model families · real-time Greeks
★ 02 · Risk & Balance Sheet

From tail risk to ALM

Continuous portfolio risk and balance sheet management on one book. Historical and parametric VaR, extreme value theory, dynamic duration gaps, and supervisory liquidity stress.

★ EVT tails · LCR & NSFR stress
★ 03 · Contagion & Networks

Systemic interconnections

Interbank network graphs, multi-round DebtRank cascades, and bilateral counterparty exposure. Built for macro supervisors and trading desks modeling cross-market contagion.

★ DebtRank · bilateral netting & XVA
★ 04 · Autonomous Desks

Research with conviction

Domain-specialized agents debate trade theses from bull, bear, and macro angles. Outputs include confidence weights, explicit invalidation triggers, and replayable reasoning traces.

★ 36 playbooks · verifiable lineage

★ 42 domains and 54 analytical hubs — unified under one typed bitemporal ontology with valid-time lineage.

★ 03   The fragmented status quo

Priced for giants,
stitched by hand for everyone else.

Funds and trading desks run on an unmaintainable patchwork: pricing from one vendor, risk from another, research in local notebooks, and execution on phone lines — stitched together with brittle Excel handoffs that break at 7am.

Their desk vs yoursIllustrative, same names
20:1
Twenty analysts on the names you cover alone.
Legacy cost floorIllustrative, incumbents
$500K+
Annual cost before the first analyst is hired.
Single-codebase hubsNyquist unified
54 hubs
One typed engine across pricing, risk, and research.
Data audit trailIllustrative, legacy
0
No valid-time lineage from filing to output when the risk committee asks why.
★ 04   The expansion trajectory

From a focused wedge
to market infrastructure.

A compounding expansion model: Nyquist lands on a single trading book, expands across multi-desk banks, and scales to sovereign market infrastructure on one shared engine. Three compounding stages of market capture.

03Market InfrastructureSIFIs & supervisors · on-prem02Multi-Desk ExpansionMulti-desk · dedicated VPC01Single-Desk WedgeFunds <$500M · cloud
01 Stage 01 · Initial WedgeLand Wedge

Boutique Funds & Desks <$500M

Replace the spreadsheet and Python glue stack for the PM who is also the risk desk. A fast 60-day paid pilot on 1–3 seats delivers immediate ROI: portfolio factor risk attribution, overnight recompute, and autonomous agent debate on live books.

WhoFunds <$500Mthe PM who is also the risk desk
List pricefrom $8,000per seat · year, Core · 1–3 seats in the pilot
Deployment PostureCloud SaaSZero-install browser app
02 Stage 02 · Enterprise ExpansionDesk Expansion

Multi-Desk Funds & Mid-Market Banks

Expands across trading and treasury desks via 54 modular hubs: Derivatives pricing, XVA, Fixed Income, balance sheet ALM, and Liquidity (LCR/NSFR). Automated risk committee packs and cross-asset netting across the organization.

WhoMulti-desk fundsregional banks & hedge funds
List pricefrom $14,500per seat · year, Professional · suites from $12,900
Deployment PostureDedicated VPCHybrid cloud or private cluster
03 Stage 03 · Market InfrastructureSovereign Scale

Tier-1 SIFIs & Central Banks

Private deployment inside the client's perimeter by contract: bitemporal audit trail, systemic risk networks, and multi-layer contagion graphs (DebtRank). Compact domain models for in-perimeter inference are on the roadmap, not in production.

WhoSIFIs & supervisorscentral banks, market authorities
Private deploymentfrom $250Kper year, by AUM band · exact terms on request
Deployment PostureOn-premclient-hosted, your GPUs
★ 05   Decision workflow

From market ticks
to an argued committee memo.

Replacing fragmented desk workflows: Nyquist computes deterministic pricing PDEs, runs adversarial bull-bear debate across named desks, and binds every conclusion to an explicit numerical kill-switch — fully audit-logged from filing to committee sign-off.

★Fig. 01 Recorded committee run · NVDA · 24 Sep 2026 · inputs as of the 23 Sep close · not investment advice.
Open the demo →
★ 06   Developers & Agents

Governed tool calling
for terminal and AI agents.

Never copy-paste risk reports into LLM prompt boxes. Nyquist operates an official Model Context Protocol (MCP) server and a lightweight terminal CLI — giving quants, DevOps engineers, and autonomous AI agents direct, verified access to deterministic PDE pricing, portfolio Greeks, and systemic stress testing.

Anthropic Open Standard

Plug the entire quantitative substrate directly into your AI agents.

No fragile copy-pasting into chat windows. Nyquist operates an official, streamable-HTTP and STDIO MCP server. When connected to Claude Desktop, Cursor, or autonomous agent runtimes, LLMs call typed mathematical and risk primitives with zero hallucinated figures.

★
58+ Governed Tools: Exposes price_bond, price_swap, compute_var,sde_calibrate_and_simulate, and run_agent_debate.
★
Lineage & Verification Block: Every tool return attaches a cryptographic digest and input parameters confirming the number came from the deterministic engine.
★
Caller-Scoped Quotas: Tokenized B2B security with isolated subaccount quotas, rate limiting, and zero customer data leakage to model vendors.
claude_desktop_config.json
{
  "mcpServers": {
    "nyquist": {
      "command": "python",
      "args": ["-m", "src.apps.mcp_node"],
      "env": {
        "NYQUIST_API_BASE": "https://api.nyquist.pro",
        "NYQUIST_MCP_API_KEY": "YOUR_MCP_TOKEN"
      }
    }
  }
}
Transport: streamable-http / stdioProtocol: 2024-11-05 (FastMCP)
Live MCP Tools:nyquist_price_swapnyquist_compute_varnyquist_run_systemic_contagionnyquist_detect_regimenyquist_run_agent_debate
★Protocol Specification · Official FastMCP streamable-HTTP & STDIO transports · Zero external runtime dependencies.Model Context Protocol spec ↗
★ 07   Field study

Everyone quotes the same vendors.
Nobody publishes the week.

We are running a field study on how risk and analytics work actually gets done — which tool the week happens in, what it costs, where it stalls. Most published research on risk workflow samples the largest institutions and models everyone else as a smaller version of them. They aren’t.

★Fig. 02 The field study in forty seconds · sixteen questions · three to five minutes.
★Sixteen questions · fourteen if you skip the early-access block · the report goes to respondents first, segment by segment, and is never resold.Take the field study →
★ 08   Start your pilot

See the platform run on your real book.

The full demo runs in your browser — no calls, no scheduling. The demo opens empty: you enter your own positions. Watch stochastic pricing, stress tests, bitemporal valid-time lineage, and autonomous research desks converge.