Chart Library MCP Server

25M+ pattern embeddings. 10 years of history. 19K+ stocks. One tool call.

Remote serverstreamable-httpPython

What is the Chart Library MCP MCP server?

Most knowledge and memory work still happens through a UI a human drives. Chart Library MCP MCP server moves it into the conversation instead. 25M+ pattern embeddings. 10 years of history. 19K+ stocks. One tool call.

The tools it exposes

The server publishes 14 tools. What each one is for:

  • search — Entry point. Find similar historical patterns for an anchor; returns a comp-set handle you can chain. mode= supports text (default), live_bars (raw
  • pull_compsThe flagship. Pull the comp set for a subject (symbol, date, timeframe) — the historical analogs, what they did next, the drivers that
  • cohort_analyze — Same engine as pull_comps under the original field names (cohort_id, feature_importance, win_rate, vol_regime, …). Kept callable verbatim for
  • cohort_introspect — Slice/probe a stored comp set by ANY attribute (macro · technical · event) and get per-subset stats vs the full-cohort baseline. No kNN re-run. *"Of
  • cohort_attribution — Within-cohort winner/loser attribution — which member traits separated the forward-return tail from the rest, each with a by-date cluster-bootstrap
  • track_record — Historical predicted-vs-realized coverage of our calibrated bands (a track record, not a forecast). The nominal 80% band held 80.8% across 302,880
  • symbol_intelligence — Layer 5 memory — per-symbol feature reliability + achieved calibration across prior analyses. Ground a read in whether a feature has historically
  • analyze — Analytic metrics. metric= accepts anomaly, volume_profile, crowding, correlation_shift, earnings_reaction, pattern_degradation, regime_accuracy
  • context — Situational data. target= accepts "market", a ticker symbol ("NVDA"), {"symbol": ..., "date": ...} for lightweight anchor metadata, or "system" for
  • explain — Narrative + rankings derived from a cohort. style= accepts filter_ranking (which filter shifts the distribution most), prose (plain-English summary)
  • portfolio — Multi-holding weighted conditional distribution. Runs per-holding cohorts in parallel, weight-averages the distributions, ranks tail contributors
  • report_feedback — File an error or improvement suggestion back to the project

Getting it running

Because this one is hosted, setup is mostly authentication — you point your client at the endpoint and approve access. Nothing runs on your machine, so there is no runtime to keep patched.

What it needs from you

Configuration is passed through the environment: CHART_LIBRARY_API_KEY. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.

How it compares

This sits in the knowledge and memory group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Chart Library MCP's toolset — search, pull_comps, cohort_analyze and 11 more — is a fair guide to whether it matches your workflow. It is maintained by grahammccain; worth a glance at recent repository activity before you build anything load-bearing on it.

We check each listing at SyncDev against the project's documentation before it goes live — if something here drifts out of date, it is a bug worth reporting.

Things to watch

  • Your data travels to the provider's service, so the usual questions apply about what you send and what they retain.
  • With 14 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Chart Library MCP.
  • Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
  • Keep per-call confirmation enabled while you learn its behaviour; it is the cheapest safeguard you have.

Available tools

ToolWhat it does
searchEntry point. Find similar historical patterns for an anchor; returns a comp-set handle you can chain. mode= supports text (default), live_bars (raw OHLCV), similar (cohort-level neighbors).
pull_comps**The flagship.** Pull the *comp set* for a subject (symbol, date, timeframe) — the historical analogs, what they did next, the **drivers** that separated the best outcomes, and our **coverage_record**. Front-of-house le
cohort_analyzeSame engine as pull_comps under the **original field names** (cohort_id, feature_importance, win_rate, vol_regime, …). Kept callable verbatim for existing integrations; new ones should prefer pull_comps.
cohort_introspectSlice/probe a stored comp set by ANY attribute (macro · technical · event) and get per-subset stats vs the full-cohort baseline. No kNN re-run. *"Of the 300 analogs, how do the post-earnings-week ones do?"*
cohort_attributionWithin-cohort winner/loser attribution — which member traits separated the forward-return tail from the rest, each with a by-date cluster-bootstrap CI and a false-discovery decision. Descriptive, never causal.
track_recordHistorical predicted-vs-realized coverage of our calibrated bands (a track record, not a forecast). The nominal 80% band held 80.8% across 302,880 prior cases.
symbol_intelligenceLayer 5 memory — per-symbol feature reliability + achieved calibration across prior analyses. Ground a read in whether a feature has historically been reliable for this ticker.
analyzeAnalytic metrics. metric= accepts anomaly, volume_profile, crowding, correlation_shift, earnings_reaction, pattern_degradation, regime_accuracy, decompose (slice winners vs losers), clusters (cohort-internal grouping).
contextSituational data. target= accepts "market", a ticker symbol ("NVDA"), {"symbol": ..., "date": ...} for lightweight anchor metadata, or "system" for DB coverage.
explainNarrative + rankings derived from a cohort. style= accepts filter_ranking (which filter shifts the distribution most), prose (plain-English summary), position_guidance (exit signals), risk_ranking.
portfolioMulti-holding weighted conditional distribution. Runs per-holding cohorts in parallel, weight-averages the distributions, ranks tail contributors.
report_feedbackFile an error or improvement suggestion back to the project.
cohort_membersThe full cohort, one record per analog, with rich per-member metadata (forward outcomes, regime, anchor fundamentals, news, chart events). Slice and bucket it yourself.
cohort_groupbyPartition the cohort by one dimension (vol_regime, sector_etf, momentum_5d, …) → per-bucket outcome distributions vs baseline. The one-call "does this dimension matter?" primitive.

How to install the Chart Library MCP MCP server

{
  "mcpServers": {
    "chart-library": {
      "command": "uvx",
      "args": ["chartlibrary-mcp"],
      "env": {
        "CHART_LIBRARY_API_KEY": "your-value"
      }
    }
  }
}

Add to claude_desktop_config.json, then restart Claude Desktop.

Configuration

VariableDescriptionRequired
CHART_LIBRARY_API_KEYCredential the server authenticates with.Yes

Example prompts to try

  • Use Chart Library MCP to search.
  • Use Chart Library MCP to pull comps.
  • Use Chart Library MCP to cohort analyze.

Frequently asked questions

It connects Chart Library MCP to MCP-compatible AI assistants such as Claude and Cursor, exposing 14 tools (search, pull_comps, cohort_analyze, and more) that the assistant can call on your behalf. Instead of copying data back and forth by hand, the assistant works with Chart Library MCP directly.