Vouch API logo

Vouch API

AI equity research that proves it isn't lying

2026-04-27

Product Introduction

  1. Definition: Vouch API is an institutional-grade financial intelligence layer and Model Context Protocol (MCP) server designed to generate and certify equity research. It functions as a "Truth Anchor" for artificial intelligence systems, providing a deterministic verification framework that bridges the gap between probabilistic Large Language Models (LLMs) and the rigorous accuracy requirements of the financial services industry.

  2. Core Value Proposition: The platform exists to eliminate financial hallucinations in AI-driven analysis, which can lead to material losses and regulatory violations. By providing cryptographic proof and source-level evidence for every claim, Vouch API enables portfolio managers, compliance teams, and hedge funds to deploy AI agents that produce defensible investment research. It integrates SEC XBRL data extraction with deterministic financial primitives to ensure that every valuation, from Discounted Cash Flow (DCF) to Monte Carlo simulations, is grounded in audited primary sources.

Main Features

  1. Deterministic Financial Primitives: Vouch API mandates the use of hard-coded computational modules rather than allowing LLMs to perform arithmetic or financial modeling through probabilistic guessing. These primitives cover essential metrics including WACC (Weighted Average Cost of Capital), NPV (Net Present Value), Terminal Value calculations, and the Treasury Stock Method for diluted shares. This ensures that the underlying math of any valuation model is 100% accurate and reproducible.

  2. Cryptographic Certification & Ed25519 Signatures: Every analysis generated by the system mints a "Decision Object." This is a tamper-evident, WORM-compliant (Write Once, Read Many) artifact signed with an Ed25519 digital signature. This certificate binds the output to its specific input data and source-level evidence, creating a permanent audit trail for regulators and internal compliance committees.

  3. SEC XBRL Deep Extraction & Mapping: The system features a sophisticated ingestion pipeline that parses structured XBRL data from 10-K, 10-Q, 8-K, and DEF 14A filings. It maps financial claims directly to us-gaap tags, providing 10-year historical time series for revenue segments, margins, and balance sheet items without manual data entry.

  4. 8-Gate Invariant Verification: Before a certificate is issued, the output must pass eight deterministic invariant checks. These include "sanity bounds" such as ensuring the fair value is within 3x of the current price, revenue growth is within 3x of the industry median, and the WACC remains within sector benchmarks. If a single gate fails, the certificate is blocked, preventing erroneous data from reaching the end-user.

  5. Model Context Protocol (MCP) & Developer Tools: Vouch API serves as an MCP server (JSON-RPC 2.0), allowing AI agents to call 12+ verified tools for financial analysis. Developers can integrate these tools into custom workflows, enabling agents to produce investment theses with verifiable citations and cryptographic hashes.

Problems Solved

  1. Pain Point: AI Hallucinations and Professional Liability. Standard LLMs often fabricate financial figures or fail at complex multi-step modeling. In a regulated environment, "guessing" leads to SEC/FINRA violations and loss of investor trust. Vouch API solves this by forcing deterministic computation and source-tracing.

  2. Target Audience: The platform is built for institutional users including Private Equity (PE) Associates, Hedge Fund Analysts, Quant Researchers, Compliance Officers (CCOs), and AI Engineers building financial agents. It specifically serves firms requiring FINRA 17a-4 and SEC 17a-4 audit trails.

  3. Use Cases:

  • CIM Diligence: Reducing the cycle for Confidential Information Memorandum (CIM) review from weeks to hours by automating data extraction and LBO modeling.
  • Automated Equity Research: Generating sell-side quality initiation reports with full evidence lineage for small-cap or non-covered stocks.
  • AI Agent Verification: Acting as a middleware safety layer for firms deploying OpenAI o-series or GPT-5 models in research workflows.

Unique Advantages

  1. Differentiation: Unlike traditional terminals like Bloomberg or FactSet which provide raw data, Vouch API certifies the reasoning and the math built on top of that data. It moves beyond "data access" to "computational certainty." While legacy systems offer partial XBRL coverage, Vouch API offers a full cryptographic provenance chain for every conclusion.

  2. Key Innovation: The "Truth Anchor" architecture. This is a unique hybrid approach that combines the reasoning capabilities of LLMs with a deterministic "referee" layer. By utilizing a gas-based economy for verifications and a reputation-based trust level system for agents, it creates an incentive structure for AI systems to remain accurate and verifiable.

Frequently Asked Questions (FAQ)

  1. How does Vouch API ensure the accuracy of AI-generated equity research? Vouch API ensures accuracy through a combination of deterministic financial primitives and 8 hard invariant gates. It prevents LLMs from performing direct calculations, instead forcing them to use pre-validated formulas for DCF, WACC, and NPV. Every output is then cross-referenced against primary SEC XBRL data and signed with an Ed25519 cryptographic signature to prove its provenance.

  2. What is a Vouch "Decision Object" and why is it important for compliance? A Decision Object is a WORM-compliant digital artifact that contains the final valuation, the evidence count, the model version, and a content hash. It is important for compliance because it meets FINRA 17a-4 and SEC 17a-4 standards for immutable record-keeping. It allows compliance teams to see exactly which filing and which line item supported a specific investment recommendation.

  3. Can Vouch API integrate with existing AI models like GPT-4 or Claude? Yes, Vouch API is built as a Model Context Protocol (MCP) server. This allows it to function as a plug-and-play toolset for any modern AI model. By connecting to the Vouch handshake endpoint, AI agents gain access to verified financial tools and can include Vouch verification hashes in their responses to prove they are not hallucinating.

  4. What financial data sources does Vouch API use for its valuations? Vouch API aggregates data from multiple authoritative sources including SEC EDGAR (XBRL structured filings), FMP, FRED (Federal Reserve Economic Data), and direct exchange feeds for real-time pricing and options flow. This "Canonical Data Spine" ensures that every number traces back to an authoritative primary source without hardcoded fallbacks.

Subscribe to Our Newsletter

Get weekly curated tool recommendations and stay updated with the latest product news