What MCP Is (and Why It Matters)
The Model Context Protocol (MCP) is an open-source standard designed to connect AI applications to external systems β data sources, tools, and workflows. [1] MCP provides a standardized way for AI applications to interact with external tools and services. [2]
The architecture is straightforward. An AI application (the MCP host) connects to one or more MCP servers, with each client maintaining a dedicated connection to its server. [3] Servers expose three core types of capabilities: Tools (executable functions), Resources (data sources), and Prompts (reusable templates). [4]
What that means practically: any AI assistant that supports MCP can call out to external services β including e-signature platforms β without you leaving the chat window.
GSD's MCP Server: What It Does
GSD has published the @gsdhere/mcp npm package, which connects MCP-compatible AI assistants to GSD's e-signature platform. The full walkthrough is on the GSD blog: gsdhere.com/blog/gsd-mcp-send-documents-from-your-ai-assistant.
The integration works with any MCP-compatible AI client. [5] From inside a chat session, you can ask your AI assistant to prepare a document for signature, review the draft, confirm the send, and track status β all without opening a separate browser tab.
Preview-Then-Confirm: The Send Never Happens Without You
One of the more important design decisions in the GSD MCP server is the preview-then-confirm flow. The AI prepares the envelope and shows you exactly what will be sent β recipient, document, any fields β before anything leaves your account. You confirm explicitly. The send does not auto-complete.
This matters because an agent calling tools without user review is a real risk in MCP-based workflows. GSD builds a two-step prepare/confirm gate directly into the server so it isn't something you have to configure or remember.
Price List Keys: The AI Can't Invent Amounts
For professionals who include pricing in their agreements β freelancers, agencies, tax preparers β there's a specific problem with letting AI draft documents: the model can hallucinate numbers. GSD's MCP server addresses this with price list keys. You define your rates in GSD; the AI references those keys rather than generating figures from context. The amounts in the signed document are the amounts you set, not what the model estimated.
This is a small feature with significant implications for anyone whose signed agreements double as billing records.
Compliance and Evidence
The signatures GSD produces are legally binding electronic signatures, ESIGN and UETA compliant in the US. Under eIDAS, GSD signatures are recognized as simple electronic signatures (SES); Advanced (AES) and Qualified (QES) signatures are available via add-ons.
Completed documents carry SHA-256 tamper-evidence. Each envelope generates a Certificate of Completion with a full audit trail, including timestamps and IP addresses. That audit trail doesn't change because the document was sent from an AI assistant rather than a web interface β the underlying signature process is the same.
Pricing
GSD operates on a straightforward tier structure:
| Plan | Price | Documents/month |
|---|---|---|
| Free | $0 | 5 |
| Pro | $12/month | β |
| Business | $49/month | β |
The Free tier is enough to evaluate the MCP integration without a credit card. Pro and Business tiers suit firms with higher volume or team needs.
Who This Is For
If you already use an AI assistant as a drafting tool, the GSD MCP server reduces the steps between drafting and getting a signature.
Ready to get signing done? Start free β 5 documents a month, no credit card β or browse the e-signature guides.