research_with_sources
Enables generating deterministic, cited research findings from caller-supplied source metadata and excerpts via a local-fixture MCP tool, supporting English and Spanish.
README
Research Scout Provider
Independent English/Spanish TypeScript provider that turns caller-supplied source metadata and excerpts into deterministic, traceable research fixtures. It exposes an HTTP API and an MCP-shaped stdio tool.
Proveedor TypeScript independiente en inglés/español que convierte metadatos y extractos proporcionados por el cliente en resultados de investigación deterministas y trazables. Expone una API HTTP y una herramienta MCP por stdio.
This service does not fetch source URLs or perform live fact-checking. It never contacts Stellar Bazaar. The public deployment is a standalone demonstration endpoint.
Live deployment: https://research-scout-provider.vercel.app
Public API
After deployment, the following routes are available:
| Method | Route | Purpose |
|---|---|---|
GET |
/ |
Provider metadata and route discovery |
GET |
/health |
Health and operating-mode status |
POST |
/v1/research |
Produce cited findings from supplied sources |
Example request:
curl -X POST "$PROVIDER_URL/v1/research" \
-H "content-type: application/json" \
-d '{
"query": "What is Stellar used for?",
"locale": "en",
"maxSources": 5,
"sources": [{
"title": "Example source",
"url": "https://example.com/article",
"excerpt": "Caller-provided evidence.",
"publisher": "Example",
"publishedAt": "2026-01-01T00:00:00Z"
}]
}'
The response contains summary, individually cited findings, normalized citations, explicit limitations, and mode: "local_fixture". Select Spanish with "locale": "es".
La respuesta incluye summary, findings con citas individuales, citations normalizadas, limitations explícitas y mode: "local_fixture". Usa "locale": "es" para español.
Input limits
- Query: 2–500 characters.
- Sources: 1–25 entries; only HTTP(S) URLs are accepted.
- Source title: up to 300 characters; excerpt: up to 2,000 characters.
maxSources: 1–10, default 5.- Local Node server request body: maximum 64 KiB.
- No credentials, wallet data, payment data, or secrets should be submitted.
Local development
Requires Node.js 20 or newer.
npm ci
npm run check
npm start
The local HTTP server listens on http://127.0.0.1:8787. To run MCP over stdio after building:
npm run build
npm run start:mcp
The MCP tool is named research_with_sources. Its input and output contract matches the HTTP endpoint.
Deploy to Vercel
This repository includes vercel.json and three serverless entry points under api/. Deploy it as a separate Vercel project:
npx vercel
npx vercel --prod
No environment variables are required. Vercel rewrites /, /health, and /v1/research to their corresponding serverless functions.
Safety and data policy
- No outbound network requests or browser automation.
- No persistence, authentication, payments, wallets, or credential handling.
- Stable source IDs are derived from source URLs with SHA-256 when omitted.
- Output synthesizes only caller-supplied text and is not independent verification.
- No integration with or publication to Stellar Bazaar.
See service-card.json for machine-readable discovery.
License
Apache License 2.0. See LICENSE.
Recommended Servers
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.
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.
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.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
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.
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.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.