tsbootstrap — Claude Code MCP server for time series
Use this MCP server to inspect time series and estimate uncertainty with bootstrap methods. It suits tasks that call for confidence intervals or forecast ranges.
claude mcp add tsbootstrap -- uvx tsbootstrapWhat tsbootstrap does
tsbootstrap gives an agent read-only tools for time-series diagnosis and confidence-interval calculations. Its library includes block and model-based resampling methods, plus reducers for panels whose series have unequal lengths. The MCP tools do not modify data.
For Claude Code, the server runs through uvx tsbootstrap; the same setup command is also listed for Codex. The library covers AR forecast intervals, EnbPI, and adaptive conformal calibration. Performance depends on the computation path: results for compiled reductions of named statistics do not describe the default NumPy backend or arbitrary Python statistics. Keep that distinction in mind when judging timings for your own workload.
What it helps with
- diagnose a time series
- calculate a bootstrap confidence interval
- estimate an AR forecast range
- reduce a panel with unequal-length series
- calibrate forecast uncertainty with EnbPI
Example prompts for the agent
- Inspect this time series and report the diagnostic findings.
- Calculate a bootstrap confidence interval for this series statistic.
- Create an AR forecast range and explain how uncertainty was estimated.
How to install tsbootstrap
Platform CLI
claude mcp add tsbootstrap -- uvx tsbootstrapuvx runs the command named after the package; if the README names another one, use uvx --from tsbootstrap <command>
Platform CLI
codex mcp add tsbootstrap -- uvx tsbootstrapuvx runs the command named after the package; if the README names another one, use uvx --from tsbootstrap <command>
Config file
{
"mcpServers": {
"tsbootstrap": {
"command": "uvx",
"args": [
"tsbootstrap"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from tsbootstrap <command>
Config file
{
"servers": {
"tsbootstrap": {
"type": "stdio",
"command": "uvx",
"args": [
"tsbootstrap"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from tsbootstrap <command>
Platform CLI
gemini mcp add tsbootstrap uvx tsbootstrapuvx runs the command named after the package; if the README names another one, use uvx --from tsbootstrap <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about tsbootstrap
How do I install tsbootstrap in Claude Code?
Run claude mcp add tsbootstrap -- uvx tsbootstrap. The server starts with uvx tsbootstrap.
What does tsbootstrap do?
It provides read-only MCP tools for time-series diagnosis and bootstrap-based interval estimates. The library also includes methods for forecast uncertainty.
Which uncertainty methods does tsbootstrap support?
The library includes block and model-based resampling, AR forecast intervals, EnbPI, and adaptive conformal calibration.
