EdgeDepth Research
MCP server for searching recorded Binance perpetual market conditions and their outcomes
- Platforms
- API, CLI
- Runs
- Locally or Remotely
- Version
- 0.11.0
- Listing updated
- today
Similar MCP servers
About EdgeDepth Research
EdgeDepth Research is an MCP server for analytics and data. It can run locally on your machine or as a hosted remote server.
The README describes EdgeDepth Research as a market microstructure search engine over recorded Binance USDT-M perpetual futures. Its MCP server lets a client such as Claude, ChatGPT, Cursor or Codex find occurrences of a defined market condition, inspect forward outcomes across the full matched set, and compare them with an unconditional reference for the same scope.
Results are returned with counts, denominators and a reproducibility key, and missing data is reported as absent rather than treated as zero. The server is positioned as research-only: the README states that no tool trades, changes alerts or publishes reports.
According to the README, connection is either hosted over Streamable HTTP with browser OAuth, or local stdio with an EdgeDepth API key. Fresh scans, cohorts and stratified computations can consume research allowance units. The package is aimed at traders and analysts who want to test a stated market claim against recorded data.
Install
Run in a terminal
claude mcp add --transport http research https://mcp.edgedepth.com/mcpAdd to Settings → Connectors → Add custom connector
https://mcp.edgedepth.com/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"research": {
"url": "https://mcp.edgedepth.com/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"research": {
"type": "http",
"url": "https://mcp.edgedepth.com/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"research": {
"serverUrl": "https://mcp.edgedepth.com/mcp"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"research": {
"type": "http",
"url": "https://mcp.edgedepth.com/mcp"
}
}
}Using several servers? Build one config file with the MCP config generator.
Configuration
EDGEDEPTH_API_KEYRequired for the stdio package. Hosted remote clients use browser OAuth instead.Required
Tools this server provides (11)
Taken from the project documentation; the running server may differ.
- list_features
- Returns the closed grammar registry of feature ids, types, operators, windows and error codes.
- list_instruments
- Returns the research universe and coverage for recorded markets.
- run_scan
- Executes a research query and returns counts, denominators, outcomes and the reproducibility key.
- next_page
- Continues a prior scan using its opaque cursor.
- run_cohort
- Compares outcomes after matched occurrences with other eligible predicate-false buckets.
- run_stratified
- Partitions a matched population into split-true, split-false and split-absent outcome summaries.
- snapshot_at
- Reads registry feature values and fired rules as of a recorded moment.
- base_rate
- Counts matches and eligible buckets for one clause over a window.
- commonality
- Finds the deterministic intersection across multiple recorded moments.
- get_report
- Retrieves a published report by its canonical hash.
- outcome_first
- Reports what preceded realised moves of a stated size, direction and horizon.
What it does
- Recorded feature search
- Queries a versioned registry of order flow, order book depth, funding, open interest, liquidations and price action features.
- Full-denominator counts
- Every result reports the eligible population and exclusions, with a reproducibility key for repeating the same query.
- Forward outcome measurement
- Computes forward returns, MFE and MAE across all occurrences of a condition, with outcome fields barred from being used as filters.
- Matched and baseline comparison
- Places the matched distribution beside other eligible predicate-false buckets in cohort results.
- Replay-linked evidence
- Representative occurrences include authenticated web links to the recorded market moment, subject to account entitlement.
- Projected large results
- Scan results are trimmed by default to keep them within client context limits, with options to restore full detail.
Good for
- Checking whether a described liquidation or order-flow condition was followed by a price move in recorded Binance perpetual data
- Counting how often a market condition occurred over a chosen window and symbol set
- Comparing outcomes after a condition against an unconditional reference for the same markets
- Opening replay links to inspect the exact recorded moment behind a result
EdgeDepth Research FAQ
Which exchange does the README say the data covers?
The README says it covers recorded Binance USDT-M crypto perpetuals.
How do I connect to the hosted server?
The README recommends connecting to https://mcp.edgedepth.com/mcp over Streamable HTTP and authorizing once in the browser, without copying an API key.
What does the local stdio package require?
The README says local stdio runs via npx with an EdgeDepth API key set as EDGEDEPTH_API_KEY and requires Node.js 20 or newer.
Can the server place trades or publish reports?
According to the README, no tool trades, modifies alerts or publishes reports.
How do I add EdgeDepth Research to Claude Code?
Run this command in your terminal: claude mcp add --transport http research https://mcp.edgedepth.com/mcp
What are the best alternatives to EdgeDepth Research?
The closest alternatives to EdgeDepth Research are MarketTrace Crypto MCP, Crypto Vertex, Deptho, Bitquery MCP Server and LiqSonar.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.
- Install steps: from the official MCP Registry entry.
