MCP and Full API

SnowCoder, inside your AI client

Drive SnowCoder's Yeti Build Agent, instance audit, and ServiceNow knowledge base from any MCP-aware AI client - no web UI required.

Available on every tier.

MCP usage draws from the same token balance as the web UI - start on Standard, scale to Enterprise when you need MSP Agents and the Yeti Build Agent.

Works with the AI clients you already use

Anything that speaks Model Context Protocol. Native support today across:

Claude Code
Claude Desktop
Cursor
Continue
GCP Vertex AI
ChatGPT Desktop
OpenAI Codex
Grok

Custom MCP clients built on the official Anthropic SDKs (TypeScript or Python) also work. Talk to Sales about any client we haven't listed.

Why MCP

Model Context Protocol (MCP) is an open standard for letting AI clients call external tools and services. Anthropic introduced MCP in 2024; in 2026 it's supported natively by Claude Desktop, Claude Code, Cursor, Continue, GCP Vertex AI Agent, OpenAI Codex, ChatGPT Desktop, and a growing list of agent frameworks.

Until now, using SnowCoder meant logging into the web UI. With MCP, SnowCoder appears as a tool inside your existing AI client. Ask Claude Desktop a ServiceNow question - Claude uses SnowCoder to answer. Ask Cursor to scaffold a scoped app - Cursor calls SnowCoder's Yeti Build Agent. Your AI client stays in the loop; SnowCoder does the ServiceNow-specific work.

What you can do with SnowCoder + MCP

Yeti-quality ServiceNow answers

Ask ServiceNow questions without leaving your AI workflow - full semantic search across our ServiceNow knowledge base.

Read and update projects, epics, stories

Drive SnowCoder's project structure directly from the AI client you already use for the rest of your work.

Generate and execute multi-artefact builds

Kick off the SnowCoder Yeti Build Agent against your target instance - 42 artifact types via the Fluent SDK with destructive-change approval gates.

Run instance audits inside an agentic workflow

Pull the audit result back into the same conversation that triggered it. 500+ granular checkpoints with severity-ranked findings.

Stay inside your existing AI workflow

No context switch into a separate web UI. No separate API plan. No separate billing - MCP usage draws from the same token balance.

Positioning

Complementary to ServiceNow Now Assist

ServiceNow shipped Now Assist for Code as part of the Australia release (May 2026). Now Assist is an in-platform AI co-pilot - useful when you're working inside the ServiceNow UI.

SnowCoder's MCP integration is out-of-platform and complementary. You drive SnowCoder from your favourite AI client (Claude Desktop, Cursor, etc.); SnowCoder does the spec-to-deploy build, audit, and verification work that Now Assist doesn't cover.

Use both: Now Assist for inline edits inside ServiceNow Studio; SnowCoder for autonomous builds from a written spec, multi-instance audits, and ServiceNow expertise in your AI client of choice.

FAQ

How do I get access?

MCP access is available on every tier. Standard customers get MCP included with their token purchases - buy tokens, generate an API key in the SnowCoder UI, and connect. Enterprise customers can layer on managed onboarding and SLA on top.

Does MCP cost extra?

No. MCP usage draws from the same token balance as the web UI. There's no separate API plan or per-call fee.

Which AI clients are supported?

Anything that speaks Model Context Protocol - Claude Code, Claude Desktop, Cursor, Continue, GCP Vertex AI, ChatGPT Desktop, OpenAI Codex, and Grok. Custom MCP clients built on the official SDKs work too.

Is it production-ready?

Yes. Live MCP server with OAuth 2.1 + PKCE + Dynamic Client Registration, refresh-token rotation, and replay-attack detection.

Are my ServiceNow credentials shared with the MCP client?

No. The MCP client only sees what SnowCoder chooses to expose. ServiceNow credentials stay encrypted in SnowCoder's database.

Does it work with my specific AI workflow?

If your client speaks MCP, yes. If you'd like us to validate compatibility with your setup before onboarding, that's part of the sales conversation.

Ready to bring SnowCoder into your AI client?

Talk to Sales for MCP access, onboarding, and a managed setup tailored to your environment.