sc-mcp
A read-only MCP server that wraps the Scalable Capital CLI to expose broker data (portfolio, holdings, transactions, analytics, quotes, news) to any MCP-capable client.
README
sc-mcp
An MCP server that wraps the
Scalable Capital sc CLI, exposing your broker
data to any MCP-capable harness (Claude Code, Claude Desktop, Codex, Cursor, …).
[!Warning] Unofficial. This is a community, read-only wrapper around Scalable Capital's official
scCLI, not affiliated with or endorsed by Scalable Capital. It's a stopgap until they ship a first-party MCP server; expect to retire it when they do.No warranty / use at your own risk. Provided "as is" under the MIT License, with no warranty of any kind. The author takes no responsibility for any loss, damage, incorrect data, or financial consequence arising from its use. This is a personal tool for your own broker data, you are responsible for verifying anything you act on. It is not financial advice.
Tools
| Tool | What it returns |
|---|---|
sc_overview |
Portfolio total value, cash, performance |
sc_holdings |
All positions with prices, quantities, market values |
sc_transactions |
Trade history with filters (date, ISIN, type, paging) |
sc_analytics |
Allocation, sector/region exposure, attribution |
sc_security_news |
Latest news summary for a security by ISIN |
sc_quote |
Current quote for a security by ISIN |
sc_search |
Search securities within the portfolio context |
sc_transaction |
Details for a single transaction by ID |
This server is read-only, it never places trades or mutates account state. All calls hit the broker live. Responses are cached in-process for 5 minutes.
The sc CLI also exposes write operations (watchlist, price-alerts,
savings-plans, trades). These are deliberately not included in this release.
Any future write support will be opt-in, disabled by default, and enabled
only via an explicit environment flag, never on by default. Money-moving
commands (trade, savings-plans) are out of scope entirely.
Prerequisites
uvPython package manager installed and available onPATH.- The
scCLI installed and onPATH. - An authenticated session:
sc login.
Compatibility
Tested against sc 0.2.x. The sc CLI is pre-1.0, so its command surface
can change between minor versions, the server logs a warning to stderr at
startup if your installed sc differs from the tested major.minor. Since sc
is an external binary, not a Python dependency, so this is the only enforcement
available. If you see the warning and a tool misbehaves, that mismatch is the
likely cause.
Install
As long as the pre-requisites are met, installing as a Claude plugin or any other agent is breezy.
Claude Code (plugin, easiest)
This repo is also a Claude Code plugin marketplace. Two commands:
/plugin marketplace add ratulotron/sc-mcp
/plugin install sc-mcp@ratulotron
That registers the scalable-capital MCP server and a usage skill. It also
bundles a light skill that tells Claude when and how to use the tools.
Claude Code (manual)
claude mcp add scalable-capital -- uvx --from git+https://github.com/ratulotron/sc-mcp@v0.1.0 sc-mcp
Cursor
Add to Cursor
— or add to .cursor/mcp.json (or ~/.cursor/mcp.json):
{
"mcpServers": {
"scalable-capital": {
"command": "uvx",
"args": ["--from", "git+https://github.com/ratulotron/sc-mcp@v0.1.0", "sc-mcp"]
}
}
}
Codex
Add to ~/.codex/config.toml:
[mcp_servers.scalable-capital]
command = "uvx"
args = ["--from", "git+https://github.com/ratulotron/sc-mcp@v0.1.0", "sc-mcp"]
Claude Desktop (bundle, no config editing)
Download sc-mcp.mcpb,
then in Claude Desktop go to Settings → Extensions → Install Extension and
pick the file.
[!Note] If Claude Desktop can't find
uvx, open the extension's settings and set the full path (e.g./opt/homebrew/bin/uvx). GUI apps on macOS don't always inherit your shellPATH.
VS Code / other MCP clients
Add the standard server config (VS Code mcp.json, or any client's MCP config):
{
"mcpServers": {
"scalable-capital": {
"command": "uvx",
"args": ["--from", "git+https://github.com/ratulotron/sc-mcp@v0.1.0", "sc-mcp"]
}
}
}
Pin the
@v0.1.0tag (see Versioning). Drop it to track the latestmain.
Configuration
| Env var | Default | Effect |
|---|---|---|
SC_MCP_CACHE_TTL |
300 |
Seconds to cache successful responses in-process. Set 0 to disable caching (always hit the broker live). |
Develop
uv sync
uv run sc-mcp # starts the stdio server
uv run pytest
Maintaining compatibility as sc evolves, when to add/update tools, bump
SUPPORTED_SC_VERSION, and the read-only invariants, is documented in
CLAUDE.md.
Versioning
This package uses SemVer. The tools are the public API:
| Bump | Trigger |
|---|---|
| MAJOR | A tool is removed/renamed, or a parameter changes incompatibly |
| MINOR | A tool or optional parameter is added |
| PATCH | Bug fix, error-message wording, internals |
Releases are tagged vX.Y.Z. Pin a tag when installing, uvx --from git+... tracks the default branch (latest) by default, so without a pin your
tool surface can change underneath you:
uvx --from git+https://github.com/ratulotron/sc-mcp@v0.1.0 sc-mcp
Most version bumps here are driven by sc CLI changes (see Compatibility), but
the version number is this package's own, it does not mirror the sc version.
License
MIT
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.