Amalgix

Amalgix

Cross-model evidence pipeline for financial filings and contracts, providing source-grounded claims with verification via MCP and REST, with x402 pay-per-call.

Category
Visit Server

README

<p align="center"> <img src="https://amalgix.io/assets/logo-icon.png" width="80" alt="Amalgix Logo"> </p>

<h1 align="center">Amalgix</h1>

<p align="center"> <strong>Cross-Model Evidence Pipeline for Financial Filings & Contracts</strong><br> <em>MCP + REST · x402 Pay-Per-Call · Up to 3.8x cheaper than calling frontier models directly</em> </p>

<p align="center"> <a href="https://amalgix.io">Website</a> | <a href="https://amalgix.io/openapi.json">OpenAPI Spec</a> | <a href="https://smithery.ai/servers/amalgix/document-intelligence">Smithery</a> | <a href="https://www.x402scan.com/server/41885358-75a2-4c56-835e-83180b18d53a">x402scan</a> | <a href="https://x.com/AmalgixHQ">X (Twitter)</a> </p>

<p align="center"> <img src="https://img.shields.io/badge/protocol-MCP-blue" alt="MCP"> <img src="https://img.shields.io/badge/payment-x402%20USDC-green" alt="x402"> <img src="https://img.shields.io/badge/chains-Base%20%7C%20Solana-purple" alt="Chains"> <img src="https://img.shields.io/badge/tools-8-orange" alt="Tools"> <img src="https://img.shields.io/badge/output-evidence--backed-brightgreen" alt="Evidence backed"> </p>


Why Amalgix?

AI agents that need evidence-backed facts from financial filings and contracts face a problem: calling frontier models like Opus 4.8 or GPT-5.5 is expensive, and a single call doesn't produce verified evidence — you need at least two passes (extract + verify) to get source-grounded output.

Amalgix solves this with a cross-model evidence pipeline:

  1. Model A extracts — A specialized extraction model pulls structured findings from your document. Each claim is tied to a verbatim quote, confidence score, and source reference.
  2. Model B verifies — A different model cross-checks every finding against the original source. Contradictions are flagged; unsupported claims are marked for review — not presented as facts.
  3. You pay less — Because each stage uses the optimal model for the job (not the most expensive one), the total pipeline cost is significantly lower than running frontier models yourself.

Cost Comparison (1MB Document)

Provider Pipeline Cost vs Amalgix
Amalgix $0.82
Opus 4.8 (2-pass) $2.88 3.5x more
GPT-5.5 (2-pass) $3.12 3.8x more

Competitor costs represent what it takes to replicate Amalgix's extract → verify output using those models directly (2-pass, ~15% + ~10% output ratio).

What is Amalgix?

A crypto-native, MCP + REST intelligence API for AI agents. The proprietary Crucible™ Evidence Engine orchestrates the cross-model pipeline:

  • Source-grounded claims — each material finding includes claim, evidence, sourceRef, confidence, and verificationStatus
  • Cross-model verification — extraction and verification use different models, catching errors that single-model approaches miss
  • Financial Filing Intelligence — SEC 10-K, 10-Q, and 20-F analysis by content, filing URL, ticker, or CIK
  • Contract Risk Intelligence — clauses, obligations, deadlines, liability, missing terms, and evidence
  • x402 payments — paid calls settle in USDC on Base or Solana. No API keys required, no subscriptions
  • Wallet dashboard — EVM wallet login for request history and API key management

Crucible is designed for evidence triage and agent workflows. It does not claim perfect accuracy, legal advice, or investment advice; responses expose confidence and verification metadata so downstream systems can inspect the basis for each finding.

Quick Start

MCP

Add Amalgix to any Streamable HTTP MCP-compatible client:

{
  "mcpServers": {
    "amalgix": {
      "url": "https://amalgix.io/mcp",
      "transport": "streamable-http"
    }
  }
}

REST

curl -X GET https://amalgix.io/api/v1/health

Paid endpoints require x402 payment headers. Use estimate_cost first to preview pricing before any paid call.

Tools

Tool Description Pricing
analyze_public_filing Financial Filing Intelligence for public SEC filings. Accepts content, filingUrl, ticker, or CIK for 10-K, 10-Q, and 20-F workflows. $0.03-$79.00
review_contract_risks Contract Risk Intelligence for clauses, obligations, liability exposure, missing clauses, unusual terms, evidence, and human-review flags. Not legal advice. $0.03-$79.00
analyze_document General evidence analysis with source-grounded findings, confidence, and verification status. $0.03-$79.00
extract_web Static/server-rendered public web content extraction with optional analysis. From $0.01
summarize Fast single-pass summarization with key points. From $0.01
delegate_bulk_translate Translate JSON string values while preserving object structure. From $0.01
estimate_cost Free pre-flight cost preview for paid tools or workflows. Free
health_check Free service health and engine availability check. Free

REST Endpoints

Endpoint Method Purpose
/api/v1/filings/analyze POST Financial Filing Intelligence
/api/v1/contracts/review POST Contract Risk Intelligence
/api/v1/analyze POST General Crucible™ evidence analysis
/api/v1/cost POST Cost estimation, free
/api/v1/health GET Health check, free
/api/v1/auth/evm/nonce POST Create wallet-login challenge, free
/api/v1/auth/evm/verify POST Verify wallet signature, free
/api/v1/api-keys GET/POST List or create API keys after wallet login, free
/api/v1/jobs GET Recent analysis jobs after wallet login, free
/api/v1/requests GET Recent request logs after wallet login, free
/api/v1/payments GET Recent x402 payment events after wallet login, free

Legacy REST tools (/api/v1/extract, /api/v1/summarize, /api/v1/translate) are available as operator opt-in via AMALGIX_ENABLE_LEGACY_REST_TOOLS=true. They remain accessible through MCP by default.

Crucible™ Output Contract

Crucible normalizes evidence into a stable shape for agents:

{
  "analysis": {
    "summary": "...",
    "evidenceFindings": [
      {
        "claim": "...",
        "evidence": "...",
        "sourceRef": "document:line:12",
        "confidence": "high",
        "verificationStatus": "supported"
      }
    ],
    "overallConfidence": "high"
  },
  "meta": {
    "engine": "Crucible™ Evidence Engine",
    "analysisType": "general_evidence",
    "routingTier": "large_context",
    "estimatedCost": 0.03
  }
}

Analysis modes:

  • general_evidence
  • financial_filing
  • contract_review

Pricing

Dynamic pricing based on document size. Amalgix's cross-model pipeline delivers equivalent evidence quality at a fraction of single-model cost:

File Size Amalgix Opus 4.8 (2-pass) GPT-5.5 (2-pass) Savings
256KB $0.30 $0.76 $0.82 59-62%
1MB $0.82 $2.88 $3.12 72-74%
5MB $7.36 $12.12 $13.14 39-44%
20MB $29.42 $46.52 $50.50 37-42%

Use the free estimate_cost tool for a request-specific quote.

Payment

Amalgix uses the x402 protocol for pay-per-call settlement:

  • Currency: USDC
  • Chains: Base (EVM) or Solana (SVM)
  • Scheme: exact
  • Authentication: x402 for paid execution; optional wallet/API key identity for dashboard and REST attribution

Security & Privacy

  • Contract raw text is not persisted by default; Amalgix stores hashes, metadata, job status, and optional results only when configured.
  • Public financial filing metadata may be cached because filings are public.
  • Connections use TLS.
  • Paid analysis content is sent to third-party LLM providers for processing.
  • Do not submit passwords, private keys, SSNs, classified data, or other regulated PII.
  • Contract and filing outputs are evidence triage only, not legal or investment advice.

Links

License

All Rights Reserved (c) 2026 Amalgix. This repository contains documentation and configuration files only. The Amalgix engine source code is proprietary and not included.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured