velora-mcp
Hosted MCP server for Argentine commerce: real AFIP/ARCA fiscal invoicing (live CAE), MercadoPago payments, logistics, catalog, cash register and WhatsApp behind one authenticated endpoint. Includes 9 no-auth fiscal validation/formatting tools.
README
This repo contains the public manifest/docs for a hosted service — the full implementation runs in production as part of Velora.
Velora — Argentine Commerce MCP
A hosted, multi-capability commerce toolkit for Argentina — real AFIP invoicing, MercadoPago, logistics, catalog and WhatsApp behind one authenticated endpoint.
Velora is a hosted Model Context Protocol server that gives AI agents tools to operate commerce in Argentina: fiscal invoicing (AFIP/ARCA), payments (MercadoPago), logistics (Andreani / PedidosYa), catalog, cash register, and WhatsApp messaging — authenticated and tenant-isolated.
Other AFIP MCP servers exist (some emit invoices, some only search the docs). Velora's angle is the bundle: fiscal is one of several commerce rails — payments, logistics, catalog, messaging — exposed through a single hosted, authenticated endpoint, so an agent runs a whole Argentine business, not just invoicing.
Endpoint
https://tools.somosvelora.com/api/mcp
Transport: streamable HTTP. Auth: API key (per-tenant HMAC) or OAuth 2.1 (WorkOS bearer). Any MCP client — Claude, OpenAI, Gemini, Cursor, or your own agent — consumes the same endpoint.
Free during early access. No usage cost while we onboard the first builders — the fiscal and pure tools run no LLM on our side, so it stays cheap to keep open.
Connect
Point any MCP client that supports a remote streamable HTTP server at the endpoint, with your credentials as headers:
{
"mcpServers": {
"velora": {
"type": "http",
"url": "https://tools.somosvelora.com/api/mcp",
"headers": {
"X-API-Key": "<your-key>",
"X-Business-Id": "<your-business-id>"
}
}
}
}
The exact config key (type/transport) varies by client — use your client's remote/HTTP MCP server format. Request access at somosvelora.com.
The pure fiscal helpers (validate_cuit, split_iva, and the rest) are also available without authentication at https://tools.somosvelora.com/api/mcp/public.
Tool packs
| Pack | What it does |
|---|---|
| Fiscal (AFIP/ARCA) | emit_invoice (real CAE — Factura A/B/C + Consumidor Final), consultar_comprobante / consultar_padron (invoice + taxpayer lookup), get_fiscal_readiness. emit_nota (NC/ND): sandbox only today — real credit/debit-note emission is in progress. |
| Pure (always on) | validate_cuit, validate_cuil, validate_cbu, build_afip_qr, format_ars, to_afip_date, split_iva, describe_invoice_type, parse_ar_amount |
| Payments | MercadoPago links, tracked payments, payment-intent status |
| Logistics | quote / create / track shipments (Andreani, PedidosYa) |
| Catalog & Stock | create / edit / delete products, stock load, bulk price update |
| Sales & Cash | register sales, returns, cash-register movements and balance |
| Customers & Suppliers | find / upsert customers, supplier management |
| Messaging | WhatsApp text & template sends |
| Reports | query_sales — sales analytics |
| Connection | connection_status — per-integration setup state (call this first) |
Maturity / what's live.
emit_invoice(Factura A/B/C + Consumidor Final), the lookup tools, and the Pure tools are production-verified against real AFIP.emit_nota(NC/ND) currently emits in sandbox only — real credit/debit-note emission is not yet wired (it returnssandbox: true, never a fake CAE). Payments, Logistics and Messaging are real tools that require per-tenant credential onboarding (connect MercadoPago / Andreani / WhatsApp) — until a tenant connects them they return demo data or fail gracefully, so don't wire them into a production flow before onboarding.emit_invoiceissues a real CAE only once ARCA onboarding is complete (check withget_fiscal_readiness); otherwise it returns a clearly-flagged sandbox response (sandbox: true), never a fake CAE presented as real.
Full machine-readable catalog — list every tool with a tools/list call against the public endpoint (no auth):
curl -s https://tools.somosvelora.com/api/mcp/public \
-H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'
Why a managed layer
AFIP's web services (WSAA + WSFE) are notoriously painful: attached PKCS#7 CMS signing, weak-DH TLS handshakes, per-CUIT certificate and token management, and a stream of new resolutions (e.g. RG 5616). Velora absorbs all of it behind clean, agent-native tool calls so you don't reimplement the fiscal stack — the same reason developers reach for managed layers over government tax APIs elsewhere in LATAM.
Status
- ✅ Real AFIP invoicing verified in production (live CAE issuance for an onboarded tenant).
- ✅ MercadoPago payments and the pure validation/formatting tools are live.
- 🔧 Logistics & messaging require per-tenant credential onboarding (see Maturity above).
Links
- Site: somosvelora.com
- Toolkit endpoint: tools.somosvelora.com
- Blog (AFIP integration field notes): somosvelora.com/blog
This repository is the public manifest and documentation for the Velora MCP server. The server itself is hosted; this repo carries the server.json registry manifest and connection docs.
Built by directing Claude Code — scoped, generated, and verified at runtime. The method is the point — the same one behind 2 fixes merged into Mastra (26k★ TS AI-agent framework): #19604, #19610.
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.