# AI Fin Hub — Full Product Notes for LLMs

## Overview
AI Fin Hub covers the intersection of AI and markets — tools, benchmarks, and long-form research.
The site publishes 33 AI-in-markets tools + 67 long-form research articles + 2 dated benchmark reports at /articles/ and /benchmarks/.

Every quantitative claim links to its source. Every tool links to its methodology. No personal trading, portfolios, or performance disclosures are referenced.

Client-side architecture: tools run in the browser; LLM-calling tools require the user to provide their own API key, which is never persisted server-side.

## Discovery Endpoints
- Homepage: https://aifinhub.io/
- AI-in-markets tools index: https://aifinhub.io/#all-tools
- Research: https://aifinhub.io/articles/
- Benchmarks: https://aifinhub.io/benchmarks/
- Methodology: https://aifinhub.io/methodology/
- Editorial standards: https://aifinhub.io/editorial-standards/
- Sponsor disclosure: https://aifinhub.io/sponsor-disclosure/
- Corrections: https://aifinhub.io/corrections/
- Agent guide: https://aifinhub.io/for-agents/
- Tool index JSON: https://aifinhub.io/agent-tools.json
- WebMCP manifest: https://aifinhub.io/.well-known/webmcp.json
- Compact LLM guide: https://aifinhub.io/llms.txt

## AI-in-Markets Tools (33 total)

> Client-side tools for AI + markets. Calculators, comparators, playgrounds, directories, generators. Browser-only; no backend. BYO API key where LLM calls are made.

### Fractional Kelly Sizer
- URL: https://aifinhub.io/kelly-sizer/
- Section: calculators
- Category: Calculators
- Summary: Map conviction tiers to fractional Kelly bet sizes with a drawdown Monte Carlo simulator. Client-side. Private by default.
- Methodology: https://aifinhub.io/methodology/kelly-sizer/

### Backtest Overfitting Score
- URL: https://aifinhub.io/backtest-overfitting-score/
- Section: calculators
- Category: Calculators
- Summary: Upload a backtest trade log and compute Probability of Backtest Overfitting (PBO), Deflated Sharpe Ratio, and the odds your edge survives live trading.
- Methodology: https://aifinhub.io/methodology/backtest-overfitting-score/

### Data-Vendor TCO Calculator
- URL: https://aifinhub.io/data-vendor-tco/
- Section: comparators
- Category: Comparators
- Summary: Compute annual cost of market data across Databento, Polygon, Alpaca, Tiingo, FMP, and Alpha Vantage for your exact universe, bar resolution, history depth, and API call volume.
- Methodology: https://aifinhub.io/methodology/data-vendor-tco/

### Finance MCP Directory
- URL: https://aifinhub.io/finance-mcp-directory/
- Section: directories
- Category: Directories
- Summary: Security-graded catalog of finance MCP servers — Alpaca, Polygon, Databento, IBKR, Tradier, Tiingo, NautilusTrader. Scope, auth, idempotency, transport, schema quality, all in one place.
- Methodology: https://aifinhub.io/methodology/finance-mcp-directory/

### Token-Cost Optimizer
- URL: https://aifinhub.io/token-cost-optimizer/
- Section: calculators
- Category: Calculators
- Summary: Compute the dollar cost of a trading research loop across Claude, GPT, and Gemini. Prompt length × model × retry × call volume → cost per idea and per validated trade.
- Methodology: https://aifinhub.io/methodology/token-cost-optimizer/

### Agent Skill Tester for Markets
- URL: https://aifinhub.io/agent-skill-tester/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste a SKILL.md definition + sample input + your Anthropic API key. See structured extraction, token cost, and latency — all in your browser. No signup, key never leaves the page.
- Methodology: https://aifinhub.io/methodology/agent-skill-tester/

### Prompt Regression Tester
- URL: https://aifinhub.io/prompt-regression-tester/
- Section: playgrounds
- Category: Playgrounds
- Summary: Run the same prompt against multiple models (Claude 4.5/4.6/4.7, GPT-5, Gemini 2.5) with your own keys. Diff outputs, score drift, catch regressions before they hit your production agent.
- Methodology: https://aifinhub.io/methodology/prompt-regression-tester/

### Hallucination Detector
- URL: https://aifinhub.io/hallucination-detector/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste a source document + an LLM's extraction. Every numeric claim in the output is checked against the source. Client-side. Catches silent fabrication before it ends up in your pipeline.
- Methodology: https://aifinhub.io/methodology/hallucination-detector/

### Kalshi vs Polymarket Arb Scanner
- URL: https://aifinhub.io/kalshi-poly-arb/
- Section: comparators
- Category: Comparators
- Summary: Daily-refreshed scan of arbitrage candidates across Kalshi and Polymarket. Paired contract matching, tax + resolution-risk overlay, no signup. Edge data rendered client-side from static JSON.
- Methodology: https://aifinhub.io/methodology/kalshi-poly-arb/

### Order Book Replay Visualizer
- URL: https://aifinhub.io/order-book-replay/
- Section: playgrounds
- Category: Playgrounds
- Summary: Drop a Level-2 CSV and watch the book reconstruct tick by tick. Animated depth bars, best bid/ask, spread over time. Understand microstructure before you design your strategy.
- Methodology: https://aifinhub.io/methodology/order-book-replay/

### Calibration Dojo
- URL: https://aifinhub.io/calibration-dojo/
- Section: playgrounds
- Category: Playgrounds
- Summary: Train your probabilistic intuition. Answer binary forecasting questions at any confidence level; track Brier score and reliability curve over time. All state stored locally.
- Methodology: https://aifinhub.io/methodology/calibration-dojo/

### Trading System Blueprinter
- URL: https://aifinhub.io/trading-system-blueprinter/
- Section: generators
- Category: Generators
- Summary: Pick your data source, LLM, broker, storage, risk engine, and logger. Get a Mermaid architecture diagram, a starter repo scaffold (ZIP), and a list of open-source integrations that actually compose.
- Methodology: https://aifinhub.io/methodology/trading-system-blueprinter/

### Risk-Adjusted Returns Calculator
- URL: https://aifinhub.io/risk-adjusted-returns/
- Section: calculators
- Category: Calculators
- Summary: Paste a returns CSV. Sharpe, Sortino, Calmar, Omega, alpha, beta, tracking error, information ratio, max drawdown, and tail moments — plus a benchmark-relative block when you include one.
- Methodology: https://aifinhub.io/methodology/risk-adjusted-returns/

### Walk-Forward Validator
- URL: https://aifinhub.io/walk-forward-validator/
- Section: playgrounds
- Category: Playgrounds
- Summary: Upload a returns CSV. Rolling or expanding IS/OOS windows, per-window Sharpe, walk-forward efficiency, and a concatenated OOS equity curve. Catches regime decay that PBO alone misses.
- Methodology: https://aifinhub.io/methodology/walk-forward-validator/

### Options Greeks Explorer
- URL: https://aifinhub.io/options-greeks-explorer/
- Section: playgrounds
- Category: Playgrounds
- Summary: Black-Scholes pricer + live Greeks visualizer. Drag spot, strike, vol, DTE, rate, dividend yield — see delta, gamma, theta, vega, rho update with the payoff curve. Call + put.
- Methodology: https://aifinhub.io/methodology/options-greeks-explorer/

### Correlation Matrix Visualizer
- URL: https://aifinhub.io/correlation-matrix-visualizer/
- Section: calculators
- Category: Calculators
- Summary: Paste a multi-asset returns CSV. See the Pearson correlation heatmap, condition number, average absolute correlation, and eigenvalue concentration — the diagnostics for detecting redundant strategies before you allocate capital.
- Methodology: https://aifinhub.io/methodology/correlation-matrix-visualizer/

### Returns Distribution Analyzer
- URL: https://aifinhub.io/returns-distribution-analyzer/
- Section: calculators
- Category: Calculators
- Summary: Paste a returns CSV. Histogram, normal-overlay, QQ plot, skewness, excess kurtosis, Jarque-Bera test, tail-weight index. See why Sharpe alone misleads when your distribution has fat tails.
- Methodology: https://aifinhub.io/methodology/returns-distribution-analyzer/

### Price-Blind Research Auditor
- URL: https://aifinhub.io/price-blind-auditor/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste a research prompt or agent context bundle. The auditor flags price numbers, directional words, and outcome-leaking phrases that cause LLMs to retroactively rationalize positions. Builds a price-blind research boundary.
- Methodology: https://aifinhub.io/methodology/price-blind-auditor/

### Prompt Injection Tester
- URL: https://aifinhub.io/prompt-injection-tester/
- Section: playgrounds
- Category: Playgrounds
- Summary: Red-team a finance agent against 24 documented prompt-injection attacks — direct override, role confusion, indirect injection via retrieved content, jailbreak patterns, tool-call hijack. BYO key; runs client-side against your live model.
- Methodology: https://aifinhub.io/methodology/prompt-injection-tester/

### Efficient Frontier Builder
- URL: https://aifinhub.io/efficient-frontier-builder/
- Section: calculators
- Category: Calculators
- Summary: Paste a multi-asset returns CSV. See the Markowitz mean-variance frontier, the minimum-variance portfolio, the max-Sharpe (tangency) portfolio, and the per-asset weights for each highlighted point. Closed-form solution, client-side.
- Methodology: https://aifinhub.io/methodology/efficient-frontier-builder/

### Options Payoff Builder
- URL: https://aifinhub.io/options-payoff-builder/
- Section: playgrounds
- Category: Playgrounds
- Summary: Build 1–4 leg option strategies. Pick call/put, long/short, strike, and contracts. See the at-expiry payoff diagram, break-even points, maximum profit and loss, and the aggregated Greeks at current spot. Presets for straddle, strangle, iron condor, spreads, butterfly.
- Methodology: https://aifinhub.io/methodology/options-payoff-builder/

### Pair Trading Cointegration Tester
- URL: https://aifinhub.io/pair-trading-tester/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste two price series. Engle-Granger cointegration test: OLS hedge ratio, Augmented Dickey-Fuller on residuals, Ornstein-Uhlenbeck half-life, z-score time series with configurable entry/exit bands. Everything client-side.
- Methodology: https://aifinhub.io/methodology/pair-trading-tester/

### Execution Simulator
- URL: https://aifinhub.io/execution-simulator/
- Section: playgrounds
- Category: Playgrounds
- Summary: Model realistic order fills — square-root market impact, linear temporary impact, latency jitter, partial fills, and queue position. See the real cost of your trade size before you route it.
- Methodology: https://aifinhub.io/methodology/execution-simulator/

### Broker API Comparator
- URL: https://aifinhub.io/broker-api-comparator/
- Section: comparators
- Category: Comparators
- Summary: Alpaca vs IBKR vs Tradier vs Schwab vs Robinhood — compare auth, rate limits, order types, market data, MCP, and fees before wiring a line of code.
- Methodology: https://aifinhub.io/methodology/broker-api-comparator/

### Synthetic Market Data Generator
- URL: https://aifinhub.io/synthetic-market-data-generator/
- Section: generators
- Category: Generators
- Summary: Generate synthetic price series — geometric Brownian motion, GARCH(1,1) with volatility clustering, regime-switching bull/bear, or copula-linked correlated pairs. Download CSV/JSON. For backtest scaffolding when you cannot share real data.
- Methodology: https://aifinhub.io/methodology/synthetic-market-data-generator/

### Financial Document Token Estimator
- URL: https://aifinhub.io/financial-document-token-estimator/
- Section: calculators
- Category: Calculators
- Summary: Paste a 10-K, 10-Q, 8-K or earnings transcript and see token count + one-pass extraction cost across eight frontier LLMs, with cache-hit toggle and context-window fit check.
- Methodology: https://aifinhub.io/methodology/financial-document-token-estimator/

### SEC Filing Chunk Optimizer
- URL: https://aifinhub.io/sec-filing-chunk-optimizer/
- Section: generators
- Category: Generators
- Summary: Pick a filing archetype, tune chunk size and overlap, and see chunk count, embedding cost, and structural-boundary warnings across three chunking strategies.
- Methodology: https://aifinhub.io/methodology/sec-filing-chunk-optimizer/

### Structured Schema Validator for Finance
- URL: https://aifinhub.io/structured-schema-validator-finance/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste LLM JSON output and validate against four pre-built finance schemas — research output, trade decision, risk snapshot, peer comparison — with sanity checks on unit and GAAP basis.
- Methodology: https://aifinhub.io/methodology/structured-schema-validator-finance/

### Agent Cost Envelope Calculator
- URL: https://aifinhub.io/agent-cost-envelope-calculator/
- Section: calculators
- Category: Calculators
- Summary: Model an LLM research loop end-to-end — steps, tool calls, convergence checks, markets per day — and see per-loop, daily, and monthly cost with cost-cap recommendations.
- Methodology: https://aifinhub.io/methodology/agent-cost-envelope-calculator/

### Fallback Chain Simulator
- URL: https://aifinhub.io/fallback-chain-simulator/
- Section: playgrounds
- Category: Playgrounds
- Summary: Define a provider fallback chain, simulate rate-limit and latency failures, and see p50/p95/p99 latency, success rate, total cost, and degradation-event distribution.
- Methodology: https://aifinhub.io/methodology/fallback-chain-simulator/

### Model Selector for Finance
- URL: https://aifinhub.io/model-selector-finance/
- Section: comparators
- Category: Comparators
- Summary: Input task, latency budget, cost budget, context size, and quality sensitivity; get ranked model recommendations with rationale — grounded in published pricing and vendor capabilities, not benchmark scores.
- Methodology: https://aifinhub.io/methodology/model-selector-finance/

### Batch vs Real-Time Cost Calculator
- URL: https://aifinhub.io/batch-vs-realtime-cost-calculator/
- Section: calculators
- Category: Calculators
- Summary: Jobs per day, tokens per job, model, deadline — get real-time vs batch cost side-by-side with savings estimate and batch-eligibility flag. Based on vendor-published batch pricing.
- Methodology: https://aifinhub.io/methodology/batch-vs-realtime-cost-calculator/

### Forecast Scoring Sandbox
- URL: https://aifinhub.io/forecast-scoring-sandbox/
- Section: playgrounds
- Category: Playgrounds
- Summary: Paste a forecast stream (probability + outcome) and see Brier score with full decomposition, log loss, reliability diagram, and bootstrap confidence intervals. Works on any prob + binary outcome CSV.
- Methodology: https://aifinhub.io/methodology/forecast-scoring-sandbox/

## Benchmarks (2 total)

> Dated, reproducible benchmark reports. Every release publishes methodology + raw data (CSV + JSON) for independent verification.

### State of AI Market Data 2026
- URL: https://aifinhub.io/benchmarks/state-of-ai-market-data-2026/
- Released: 2026-04-20 · Q2 release
- Description: 2026 Q2 baseline for AI-in-markets infrastructure: vendor pricing, MCP grades, LLM per-task cost.
- Raw data CSV: https://aifinhub.io/benchmarks/state-of-ai-market-data-2026/data.csv
- Raw data JSON: https://aifinhub.io/benchmarks/state-of-ai-market-data-2026/data.json

### State of LLM Pricing for Finance Q2 2026
- URL: https://aifinhub.io/benchmarks/state-of-llm-pricing-for-finance-q2-2026/
- Released: 2026-04-23 · Q2 pricing snapshot
- Description: Published vendor pricing for 10 frontier LLMs — Claude, GPT-5, Gemini — with cache, batch, thinking-token, and context-window axes. Four worked cost scenarios.
- Raw data CSV: https://aifinhub.io/benchmarks/state-of-llm-pricing-for-finance-q2-2026/data.csv
- Raw data JSON: https://aifinhub.io/benchmarks/state-of-llm-pricing-for-finance-q2-2026/data.json

## Research Articles (67 total)

> Long-form research on LLMs in markets. Pillar guides, head-to-head comparisons, runnable tutorials, methodology pieces. Education only, not investment advice.

### Model Selection for Finance Teams: A Methodology That Survives Bench Cycles
- URL: https://aifinhub.io/articles/model-selection-finance-survives-bench-cycles/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Model selection finance methodology: a five-axis rubric, quarterly rebench cadence, version-pinning, and shadow A/B that survive 3-6 month frontier-release churn.
- Reading time: 13 min
- Tools referenced: model-selector-finance, prompt-regression-tester, agent-skill-tester, token-cost-optimizer, fallback-chain-simulator

### Backtest Overfitting in LLM Trading Strategies: PBO Score Explained
- URL: https://aifinhub.io/articles/backtest-overfitting-llm-strategies-pbo-explained/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: The Probability of Backtest Overfitting, applied to LLM-augmented research. Why LLM strategies inflate PBO, how to compute it, and the three-gate validation flow.
- Reading time: 13 min
- Tools referenced: backtest-overfitting-score, walk-forward-validator, calibration-dojo

### Caching Strategies for Production LLM Pipelines: A 2026 Cost Analysis
- URL: https://aifinhub.io/articles/caching-strategies-llm-pipelines-2026/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Three caching layers for production LLM pipelines: provider-side prompt caching, application response cache, semantic similarity cache. Decision matrix + worked numbers.
- Reading time: 11 min
- Tools referenced: token-cost-optimizer, agent-cost-envelope-calculator, data-vendor-tco

### Hallucination Detection at Scale: From R&D to Production Pipelines
- URL: https://aifinhub.io/articles/hallucination-detection-at-scale-production/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Production-grade LLM hallucination detection in four layers: source grounding, self-consistency, deterministic verification, and adversarial probes. With a worked 10-Q research agent.
- Reading time: 12 min
- Tools referenced: hallucination-detector, agent-skill-tester, prompt-regression-tester, price-blind-auditor

### MCP Servers for Financial Data: A Security-Graded Catalog Walkthrough
- URL: https://aifinhub.io/articles/mcp-servers-financial-data-security-graded/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: A five-grade A-E rubric for finance MCP servers across auth, egress, audit, rotation, and vendor posture. 12-server walkthrough plus anti-patterns degrading B-grade installs to D-grade reality.
- Reading time: 13 min
- Tools referenced: finance-mcp-directory, data-vendor-tco, structured-schema-validator-finance

### Production LLM Latency Budgets: P50/P95/P99 Math for Trading Apps
- URL: https://aifinhub.io/articles/production-llm-latency-budgets-trading/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Trading apps with LLM-augmented research need explicit latency budgets per call. The P50/P95/P99 math, queue-theory bounds, and architecture patterns.
- Reading time: 11 min
- Tools referenced: agent-cost-envelope-calculator, execution-simulator

### Token-Cost Optimization: Prompt Caching vs Distillation vs Retrieval
- URL: https://aifinhub.io/articles/token-cost-prompt-cache-vs-distill-vs-rag/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Three token-cost reduction strategies. Decision rules for when each pays back, and the math when they compound.
- Reading time: 11 min
- Tools referenced: token-cost-optimizer, agent-cost-envelope-calculator

### Compliance Audit Trails for LLM-Driven Trade Decisions
- URL: https://aifinhub.io/articles/compliance-audit-trails-llm-trades/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Schema, append-only log design, and reproducibility patterns for SEC/FINRA-compliant LLM trade audit trails.
- Reading time: 12 min
- Tools referenced: hallucination-detector, prompt-regression-tester

### Vendor Lock-In Risk: How to Architect Cross-Provider Fallback Chains
- URL: https://aifinhub.io/articles/vendor-lock-in-cross-provider-fallback/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Anthropic, OpenAI, and Google can all break or price-jump in one quarter. The fallback-chain architecture that survives a single-vendor outage.
- Reading time: 11 min
- Tools referenced: model-selector-finance, fallback-chain-simulator

### The Retrieval-Augmented Generation Cost Model: When RAG Loses to Fine-Tuning
- URL: https://aifinhub.io/articles/rag-cost-model-vs-fine-tuning/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: RAG looks cheaper at low query volume. Above 100k queries/month with stable knowledge, fine-tuning wins. The break-even math.
- Reading time: 11 min
- Tools referenced: agent-cost-envelope-calculator, token-cost-optimizer

### Prompt Version Control: Git Workflows for LLM Engineering Teams
- URL: https://aifinhub.io/articles/prompt-version-control-git-workflows/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Prompts are code. Version them, diff them, regression-test them. Three workflow patterns that survive team scaling.
- Reading time: 10 min
- Tools referenced: prompt-regression-tester, agent-skill-tester

### MCP Server Latency: The Hidden Cost of Tool-Call Roundtrips
- URL: https://aifinhub.io/articles/mcp-server-latency-tool-call-roundtrips/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Each MCP tool call adds latency. For multi-step agents, the total roundtrip cost dominates. Architecture patterns to amortize it.
- Reading time: 10 min
- Tools referenced: finance-mcp-directory, agent-skill-tester

### Cost-Per-Validated-Trade: A Framework for LLM ROI in Trading
- URL: https://aifinhub.io/articles/cost-per-validated-trade-framework/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Tokens spent doesn't equal value. The cost-per-validated-trade metric and how to instrument it.
- Reading time: 11 min
- Tools referenced: token-cost-optimizer, calibration-dojo, kelly-sizer

### Calibration Drift: Why Your LLM's Confidence Score Lies After 3 Months
- URL: https://aifinhub.io/articles/calibration-drift-llm-confidence-scores/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: LLM-reported confidence calibration drifts as the model is updated. Detection patterns and re-calibration math.
- Reading time: 11 min
- Tools referenced: calibration-dojo, prompt-regression-tester

### Temperature, Top-P, and Top-K: When Each Costs You Money in Production
- URL: https://aifinhub.io/articles/temperature-top-p-top-k-production-cost/
- Archetype: Pillar · Guide
- Published: 2026-05-07
- Description: Sampling parameters affect both quality and cost. The decision rules for each parameter, with worked examples on production cost impact.
- Reading time: 10 min
- Tools referenced: token-cost-optimizer, agent-skill-tester

### Reading Financial Filings With LLMs: 2026 Playbook
- URL: https://aifinhub.io/articles/reading-financial-filings-with-llms-2026/
- Archetype: Pillar · Guide
- Published: 2026-04-23
- Description: A map of eight filing tasks — extraction, summarization, peer comparison, Q&A, classification, sentiment, forecasting input, compliance — with model, pattern, and cost.
- Reading time: 14 min
- Tools referenced: financial-document-token-estimator, sec-filing-chunk-optimizer, structured-schema-validator-finance, hallucination-detector, token-cost-optimizer

### Fine-Tuning vs RAG vs Long-Context for Filings
- URL: https://aifinhub.io/articles/finetune-vs-rag-vs-longcontext-filings/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: Decision matrix for finance LLMs: when RAG wins, when long-context wins, and when fine-tuning makes sense. Cost math from published 2026-04 vendor rates.
- Reading time: 12 min
- Tools referenced: financial-document-token-estimator, sec-filing-chunk-optimizer, token-cost-optimizer

### Prompt Caching Economics for Finance
- URL: https://aifinhub.io/articles/prompt-caching-economics-finance/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: How Anthropic, OpenAI, and Gemini prompt caching works on finance workloads — 5-minute TTL, hit-rate patterns, and 50-90% input savings at the right design.
- Reading time: 11 min
- Tools referenced: financial-document-token-estimator, token-cost-optimizer, batch-vs-realtime-cost-calculator

### Numeric Precision in LLM Filing Extraction
- URL: https://aifinhub.io/articles/numeric-precision-llm-filings/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Six precision traps — units, currency, GAAP vs non-GAAP, diluted vs basic shares, restatements, rounding — and the structured-output pattern that fixes them.
- Reading time: 11 min
- Tools referenced: structured-schema-validator-finance, hallucination-detector, agent-skill-tester

### Prompt Patterns for Earnings Calls
- URL: https://aifinhub.io/articles/prompt-patterns-earnings-calls/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Five copy-paste patterns — speaker attribution, hedged-guidance confidence, multi-quarter delta, risk aggregator, forward-outlook separator — with runnable code.
- Reading time: 12 min
- Tools referenced: agent-skill-tester, prompt-regression-tester, structured-schema-validator-finance

### Observability Patterns for LLM Trading Agents
- URL: https://aifinhub.io/articles/observability-llm-trading-agents/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Three patterns that stop silent failure: trace-ID propagation, structured log schema with per-step cost and confidence, and a deterministic replay harness.
- Reading time: 11 min
- Tools referenced: trading-system-blueprinter, agent-cost-envelope-calculator, token-cost-optimizer

### MCP vs Function Calling for Finance Agents
- URL: https://aifinhub.io/articles/mcp-vs-function-calling-finance/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: Where MCP wins, where function calling wins, and why the right answer is almost always a hybrid — data layer on MCP, decision code on function calling.
- Reading time: 10 min
- Tools referenced: finance-mcp-directory, agent-skill-tester, trading-system-blueprinter

### Rate Limit Design for LLM Research Loops
- URL: https://aifinhub.io/articles/rate-limit-design-llm-research/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Three primitives that turn bursty finance workloads into stable loops: per-provider token bucket, cross-provider fallback chain, and graceful degradation.
- Reading time: 10 min
- Tools referenced: fallback-chain-simulator, token-cost-optimizer, trading-system-blueprinter

### Bounded-Cost Agentic Research
- URL: https://aifinhub.io/articles/bounded-cost-agentic-research/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: Three gates stop runaway agent loops: hard token budget, step-count cap, and a cost-convergence check that halts when belief stops moving.
- Reading time: 10 min
- Tools referenced: agent-cost-envelope-calculator, token-cost-optimizer, trading-system-blueprinter

### Prompt Injection Defenses for Finance Agents
- URL: https://aifinhub.io/articles/prompt-injection-defenses-finance/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: Five stacked defenses: input fencing, output validation, tool allow-list, bounded-cost circuit, dual-model cross-check. No single defense is sufficient.
- Reading time: 11 min
- Tools referenced: prompt-injection-tester, hallucination-detector, agent-skill-tester

### Agent Memory Patterns for Finance Research
- URL: https://aifinhub.io/articles/agent-memory-patterns-finance/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Three memory tiers for finance agents — working, episodic, long-term lesson library — with retention policies and runnable Python for each.
- Reading time: 11 min
- Tools referenced: agent-skill-tester, agent-cost-envelope-calculator, trading-system-blueprinter

### Model Selection Framework for Finance Tasks
- URL: https://aifinhub.io/articles/model-selection-framework-finance/
- Archetype: Comparison · Benchmark
- Published: 2026-04-23
- Description: A task × latency × cost × context decision tree for finance LLM workloads. Ten concrete scenarios mapped to tier bands. Grounded in published pricing, not benchmarks.
- Reading time: 12 min
- Tools referenced: model-selector-finance, token-cost-optimizer, financial-document-token-estimator

### Batch API Economics for Finance Loops
- URL: https://aifinhub.io/articles/batch-api-economics-finance/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: When Anthropic Message Batches or OpenAI Batch cut cost by half on finance workloads — and the soft-deadline rule for when batch is not a valid choice.
- Reading time: 11 min
- Tools referenced: batch-vs-realtime-cost-calculator, financial-document-token-estimator, token-cost-optimizer

### Thinking Tokens for Finance Tasks
- URL: https://aifinhub.io/articles/thinking-tokens-finance-tasks/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: When extended-thinking and reasoning-effort modes earn their 3-10x cost tax on finance workloads — and when they are a silent drain on the budget.
- Reading time: 11 min
- Tools referenced: model-selector-finance, token-cost-optimizer, financial-document-token-estimator

### Inference Cost Attribution per Idea and Trade
- URL: https://aifinhub.io/articles/inference-cost-attribution-trade/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Append-only cost-event schema plus two canonical SQL queries — cost per idea, cost per validated trade — with cache-write amortization built in.
- Reading time: 11 min
- Tools referenced: agent-cost-envelope-calculator, token-cost-optimizer, batch-vs-realtime-cost-calculator

### Context Hygiene for Multi-Step Research
- URL: https://aifinhub.io/articles/context-hygiene-multi-step-research/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Three-tier layered summary — leaf documents, intermediate briefs, working memory — with per-tier retention rules that keep long research loops cheap and sharp.
- Reading time: 10 min
- Tools referenced: financial-document-token-estimator, token-cost-optimizer, agent-skill-tester

### Evaluation Harness for Finance LLM Tasks
- URL: https://aifinhub.io/articles/eval-harness-finance-llm/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Why public benchmarks are a signal not a decision, how to source ground truth from EDGAR, and a runnable eval-harness skeleton with bootstrap confidence intervals.
- Reading time: 12 min
- Tools referenced: agent-skill-tester, prompt-regression-tester, hallucination-detector

### Bayesian Updating for LLM-Assisted Forecasts
- URL: https://aifinhub.io/articles/bayesian-updating-llm-forecasts/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Turn LLM probability outputs into calibrated posteriors — Beta-Binomial for binary forecasts, Normal-Inverse-Gamma for continuous — with runnable Python.
- Reading time: 11 min
- Tools referenced: calibration-dojo, kelly-sizer, backtest-overfitting-score

### Brier Scores and Log Loss for Forecasters
- URL: https://aifinhub.io/articles/brier-scores-log-loss-forecasters/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Two proper scoring rules for probabilistic forecasts, why Brier decomposes into reliability plus resolution, and why log loss punishes overconfident wrongness.
- Reading time: 10 min
- Tools referenced: forecast-scoring-sandbox, calibration-dojo, kelly-sizer

### Research Diary Schema: Auditable LLM Research
- URL: https://aifinhub.io/articles/research-diary-schema-auditable/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: A 12-field append-only schema that captures every idea — including rejected ones — to unlock calibration, proper scoring, and post-hoc overfitting analysis.
- Reading time: 10 min
- Tools referenced: calibration-dojo, backtest-overfitting-score, forecast-scoring-sandbox

### Multi-Timeframe Signal Integration With LLMs
- URL: https://aifinhub.io/articles/multi-timeframe-signal-integration-llms/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: LLMs belong on weekly fundamentals, not intraday microstructure. A two-layer architecture: weekly LLM thesis plus rule-based intraday invalidation gates.
- Reading time: 11 min
- Tools referenced: correlation-matrix-visualizer, kelly-sizer, token-cost-optimizer

### News Feed Integration for Finance Agents
- URL: https://aifinhub.io/articles/news-feed-integration-finance-agents/
- Archetype: Tutorial · Runnable
- Published: 2026-04-23
- Description: Four patterns — source vetting, injection sanitization, timestamp discipline, dedup across reporters — make news safe for an LLM finance agent. Runnable scaffold.
- Reading time: 11 min
- Tools referenced: prompt-injection-tester, hallucination-detector, agent-skill-tester

### After-Hours, 24-7, and Pre-Market Asymmetries
- URL: https://aifinhub.io/articles/after-hours-247-premarket-asymmetries/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: Three boundaries where LLM research built on equity's 9:30–16:00 clock breaks — earnings after close, 24-7 crypto, pre-market Asia/Europe action. Decision rule.
- Reading time: 10 min
- Tools referenced: execution-simulator, token-cost-optimizer, returns-distribution-analyzer

### Postmortem Template for LLM Trading Systems
- URL: https://aifinhub.io/articles/postmortem-template-llm-trading-systems/
- Archetype: Methodology · Opinion
- Published: 2026-04-23
- Description: A blameless, append-only postmortem template plus a 20-mode failure checklist — price-blind leaks to cache poisoning — keyed to the trace-ID log.
- Reading time: 10 min
- Tools referenced: price-blind-auditor, hallucination-detector, prompt-injection-tester

### Walk-Forward Validation: A Cookbook
- URL: https://aifinhub.io/articles/walk-forward-validation-cookbook/
- Archetype: Tutorial · Runnable
- Published: 2026-04-22
- Description: Walk-forward is the cheapest honest backtest you can run. Anchored vs rolling windows, the four parameters that matter, and a 60-line Python template.
- Reading time: 11 min
- Tools referenced: walk-forward-validator, backtest-overfitting-score, risk-adjusted-returns

### Prompt Injection Attack Catalog for Finance Agents
- URL: https://aifinhub.io/articles/prompt-injection-attack-catalog-finance/
- Archetype: Methodology · Opinion
- Published: 2026-04-22
- Description: Prompt injection attacks on finance agents — indirect injection via news feeds, tool-result poisoning, prompt exfiltration, unit confusion — plus defenses.
- Reading time: 10 min
- Tools referenced: prompt-injection-tester, hallucination-detector, price-blind-auditor

### Backtest to Paper to Live: Deployment Playbook
- URL: https://aifinhub.io/articles/backtest-to-paper-to-live-playbook/
- Archetype: Tutorial · Runnable
- Published: 2026-04-22
- Description: Backtest to paper to live — the gates that separate each stage, the metrics that trigger rollback, and the kill-switch you should already have.
- Reading time: 12 min
- Tools referenced: trading-system-blueprinter, execution-simulator, backtest-overfitting-score

### The $0/Month Trading Stack in 2026
- URL: https://aifinhub.io/articles/zero-dollar-trading-stack/
- Archetype: Methodology · Opinion
- Published: 2026-04-22
- Description: Zero-cost solo trading stack: launchd + free market data tiers + local LLMs on cheap paths + BYO API keys — plus where paid tiers become unavoidable.
- Reading time: 10 min
- Tools referenced: data-vendor-tco, token-cost-optimizer, trading-system-blueprinter

### Real-Time vs End-of-Day Trading Systems
- URL: https://aifinhub.io/articles/real-time-vs-end-of-day-systems/
- Archetype: Methodology · Opinion
- Published: 2026-04-22
- Description: Real-time vs end-of-day trading systems — the decision rule, the 20-50x cost delta, and the four signal types where real-time is genuinely load-bearing.
- Reading time: 9 min
- Tools referenced: data-vendor-tco, execution-simulator

### Rate-Limited, Resumable Market-Data Ingestion
- URL: https://aifinhub.io/articles/rate-limited-resumable-market-data-ingestion/
- Archetype: Tutorial · Runnable
- Published: 2026-04-22
- Description: Four primitives that turn a weekend ingestion script into a six-month loop: token-bucket limits, resumable checkpoints, idempotent writes, DLQs.
- Reading time: 11 min
- Tools referenced: data-vendor-tco, trading-system-blueprinter

### Synthetic Market Data for Backtests: Beyond GBM
- URL: https://aifinhub.io/articles/synthetic-market-data-for-backtests/
- Archetype: Tutorial · Runnable
- Published: 2026-04-22
- Description: Synthetic market data beyond GBM — when GARCH(1,1), regime-switching, or copula-linked pairs are the right next step. Trade-offs plus a Python template.
- Reading time: 10 min
- Tools referenced: synthetic-market-data-generator, backtest-overfitting-score, walk-forward-validator

### Execution Simulation: Slippage and Impact
- URL: https://aifinhub.io/articles/execution-simulation-slippage-impact/
- Archetype: Methodology · Opinion
- Published: 2026-04-22
- Description: The math of market impact — why it scales as the square root of trade size, when linear impact dominates, and the fix that keeps backtests honest.
- Reading time: 11 min
- Tools referenced: execution-simulator, order-book-replay, risk-adjusted-returns

### Choosing a Broker API in 2026
- URL: https://aifinhub.io/articles/choosing-a-broker-api-2026/
- Archetype: Comparison · Benchmark
- Published: 2026-04-22
- Description: Choosing a broker API 2026 — Alpaca vs IBKR vs Tradier vs Schwab vs Robinhood on the axes that bite: auth, order types, rate limits, and fees.
- Reading time: 10 min
- Tools referenced: broker-api-comparator, finance-mcp-directory, data-vendor-tco

### Options Greeks for LLM-Driven Trading
- URL: https://aifinhub.io/articles/options-greeks-for-llm-driven-trading/
- Archetype: Tutorial · Runnable
- Published: 2026-04-22
- Description: Options Greeks for LLM-driven trading: delta, gamma, theta, vega, rho — what each costs, three rules, plus a prompt template for multi-leg positions.
- Reading time: 10 min
- Tools referenced: options-greeks-explorer, options-payoff-builder, agent-skill-tester

### The Sharpe Ratio Trap
- URL: https://aifinhub.io/articles/sharpe-ratio-trap/
- Archetype: Methodology · Opinion
- Published: 2026-04-22
- Description: Sharpe ignores tail risk, assumes Gaussian returns, and is trivially gameable. Four metrics to report alongside it: Sortino, Calmar, tail, deflated Sharpe.
- Reading time: 8 min
- Tools referenced: risk-adjusted-returns, returns-distribution-analyzer, backtest-overfitting-score

### The 2026 Engineer's Guide to AI in Markets
- URL: https://aifinhub.io/articles/the-2026-engineers-guide-to-ai-in-markets/
- Archetype: Pillar · Guide
- Published: 2026-04-20
- Description: An engineer's map of where LLMs, MCP servers, and market-data APIs fit into a 2026 trading stack — and where they still break. Direct, no hype, no grift.
- Reading time: 10 min
- Tools referenced: data-vendor-tco, finance-mcp-directory, token-cost-optimizer, backtest-overfitting-score, kelly-sizer

### Market Data APIs Compared: Databento vs Polygon 2026
- URL: https://aifinhub.io/articles/market-data-apis-compared-2026/
- Archetype: Comparison · Benchmark
- Published: 2026-04-20
- Description: Market data APIs compared: six retail providers on pricing, tier coverage, real-time access, options and futures coverage, and who wins for each profile.
- Reading time: 9 min
- Tools referenced: data-vendor-tco, finance-mcp-directory

### Broker APIs Compared: Alpaca vs IBKR vs Tradier 2026
- URL: https://aifinhub.io/articles/broker-apis-compared-2026/
- Archetype: Comparison · Benchmark
- Published: 2026-04-20
- Description: Broker APIs for retail AI trading 2026: Alpaca for solo ops (official MCP), IBKR for multi-asset depth, Tradier for options. Head-to-head + MCP angle.
- Reading time: 8 min
- Tools referenced: finance-mcp-directory, data-vendor-tco, trading-system-blueprinter

### Finance MCP Servers: The Security Baseline
- URL: https://aifinhub.io/articles/finance-mcp-security-baseline/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: An opinionated rubric for grading 2026 finance MCP servers on scope, auth, idempotency, transport, and schema — plus the failure modes that kill agents.
- Reading time: 11 min
- Tools referenced: finance-mcp-directory, trading-system-blueprinter

### The Price-Blind LLM Research Harness
- URL: https://aifinhub.io/articles/price-blind-llm-research-harness/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: Price-blind LLM research — most harnesses leak the current price and the model confabulates. The architectural fix and a 30-line Python scaffold.
- Reading time: 8 min
- Tools referenced: prompt-regression-tester, hallucination-detector, calibration-dojo

### The 8-Step LLM Research Prompt Template
- URL: https://aifinhub.io/articles/8-step-llm-research-prompt-template/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: Free-form prompts yield uncalibrated output. An 8-step template — reference class, decomposition, pre-mortem, invalidation, JSON — fixes that.
- Reading time: 9 min
- Tools referenced: prompt-regression-tester, agent-skill-tester

### LLM Prompt Patterns for 10-K and 8-K Extraction
- URL: https://aifinhub.io/articles/llm-prompt-patterns-10k-extraction/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: Three structured patterns for auditable 10-K extractions: field-by-field JSON, citation-required verbatim quotes, and contradiction-triangle cross-check.
- Reading time: 9 min
- Tools referenced: hallucination-detector, agent-skill-tester, token-cost-optimizer

### Did You Overfit? PBO and Deflated Sharpe
- URL: https://aifinhub.io/articles/did-you-overfit-pbo-deflated-sharpe/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: A practical tutorial on the two best-documented tests for backtest overfitting — PBO via CSCV and the Deflated Sharpe Ratio. Runnable Python + tool.
- Reading time: 12 min
- Tools referenced: backtest-overfitting-score, kelly-sizer

### Signal Orthogonality: Why Ensembles Become One Bet
- URL: https://aifinhub.io/articles/signal-orthogonality-ensembles/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: A 10-signal ensemble with pairwise correlation 0.8 is effectively a 1.5-signal ensemble. The math, a two-minute diagnostic, and three axes that work.
- Reading time: 7 min
- Tools referenced: backtest-overfitting-score, kelly-sizer

### Conviction-Scaled Kelly Bet Sizing
- URL: https://aifinhub.io/articles/conviction-scaled-kelly/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: Full Kelly is brutally unforgiving of over-estimation. Quarter-Kelly with a conviction-tier mapping and a per-trade cap is the defensible default.
- Reading time: 8 min
- Tools referenced: kelly-sizer, backtest-overfitting-score, calibration-dojo

### Calibrating LLM Forecasts with Isotonic Regression
- URL: https://aifinhub.io/articles/isotonic-calibration-llm-forecasts/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: LLM probabilities are systematically miscalibrated. Isotonic regression via PAV is the cheapest robust fix: 40 lines of Python, no distributional priors.
- Reading time: 10 min
- Tools referenced: calibration-dojo, kelly-sizer, backtest-overfitting-score

### Heartbeats, Watchdogs, Circuit Breakers for Trading
- URL: https://aifinhub.io/articles/heartbeats-watchdogs-circuit-breakers/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: Silent failure is the worst failure mode. Three patterns prevent it — heartbeat, watchdog, circuit breaker — in under 100 lines of Python on launchd.
- Reading time: 9 min
- Tools referenced: trading-system-blueprinter, finance-mcp-directory, token-cost-optimizer

### The Token-Cost Reality of LLM Trading Research
- URL: https://aifinhub.io/articles/token-cost-reality-llm-trading-research/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: What LLM trading research costs per idea and per validated trade across Claude, GPT-5, and Gemini 2.5. Pricing, caching, model-mix under $200/month.
- Reading time: 8 min
- Tools referenced: token-cost-optimizer, prompt-regression-tester

### Building a Production Claude Agent for Finance
- URL: https://aifinhub.io/articles/production-claude-agent-for-finance/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: Production Claude agent for finance: price-blind research, idempotent execution, heartbeat + watchdog + circuit breaker, under $225/month at small scale.
- Reading time: 14 min
- Tools referenced: trading-system-blueprinter, token-cost-optimizer, kelly-sizer, finance-mcp-directory, walk-forward-validator

### The BaFin + EU Guide for Retail AI Traders (2026)
- URL: https://aifinhub.io/articles/bafin-eu-guide-retail-ai-traders/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: BaFin and EU rules for retail AI trading, publishing finance content, and automated strategies. Education-safe phrasing and the minimum compliance stack.
- Reading time: 12 min

### The 5 Failure Modes of LLM Trading Agents (2026)
- URL: https://aifinhub.io/articles/5-failure-modes-llm-trading-agents/
- Archetype: Methodology · Opinion
- Published: 2026-04-20
- Description: The 5 recurring failure modes in retail LLM trading agents: price-blind leaks, numeric fabrication, prompt drift, token runaway, audit amnesia.
- Reading time: 8 min
- Tools referenced: price-blind-auditor, hallucination-detector, prompt-regression-tester, token-cost-optimizer, trading-system-blueprinter

### How to Read a Backtest Report: 2026 Cheat Sheet
- URL: https://aifinhub.io/articles/how-to-read-a-backtest-report/
- Archetype: Tutorial · Runnable
- Published: 2026-04-20
- Description: Five questions a backtest report must answer — edge real, persistent, cheap to trade, bearable, explainable — with the statistics that verify each.
- Reading time: 9 min
- Tools referenced: backtest-overfitting-score, walk-forward-validator, risk-adjusted-returns, returns-distribution-analyzer, correlation-matrix-visualizer

## Resources (0 total)