Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Polymarket MCP Bot Analyst Show me the top 10 traders in the last 7 days"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for analyzing successful trading bots on Polymarket ā discover top traders, classify their strategies with AI, and detect bots on the world's largest prediction market.
Related MCP server: polymarket-mcp
This project implements a Model Context Protocol (MCP) server that exposes three powerful tools for analyzing trading activity on Polymarket. It combines real-time leaderboard data from Polymarket's Data API with LLM-powered strategy classification via OpenAI.
š Top Trader Discovery ā Fetch leaderboard rankings by timeframe
š§ AI Strategy Analysis ā Classify strategies (arbitrage, market-making, etc.) using GPT-4o-mini
š¤ Bot Detection ā Heuristic + LLM-based identification of automated traders
š Batch Reporting ā Concurrent analysis of multiple profiles
š Resilient API Layer ā Exponential backoff, rate-limit handling (429 + Retry-After), graceful fallbacks
graph TD
subgraph Client ["Client Layer"]
MCP_Client["MCP Client (e.g., Claude Desktop)"]
end
subgraph Server ["MCP Server Layer"]
index["index.ts (McpServer)"]
Validation["Zod Validation"]
end
subgraph Tools ["Tool Handlers"]
Traders["traders.ts (find_top_traders)"]
Analysis["analysis.ts (analyze_trader_strategy)"]
Reports["reports.ts (generate_batch_report)"]
end
subgraph Services ["External Services & Utils"]
PAPI["api/polymarket.ts (Polymarket Data API)"]
LLM["utils/llm.ts (OpenAI GPT-4o-mini)"]
end
MCP_Client -- "stdio (JSON-RPC)" --> index
index --> Validation
Validation --> Traders
Validation --> Analysis
Validation --> Reports
Traders --> PAPI
Analysis --> PAPI
Analysis --> LLM
Reports --> Analysis
Reports --> PAPI
style Client fill:#f9f,stroke:#333,stroke-width:2px
style Server fill:#bbf,stroke:#333,stroke-width:2px
style Tools fill:#dfd,stroke:#333,stroke-width:2px
style Services fill:#ffd,stroke:#333,stroke-width:2pxsequenceDiagram
participant C as MCP Client
participant S as MCP Server
participant T as Tool Handler
participant P as Polymarket API
participant L as OpenAI LLM
C->>S: Call "analyze_trader_strategy"
S->>S: Validate Input (Zod)
S->>T: handleAnalyzeStrategy(profile_id)
T->>P: Fetch Profile Data & PnL
P-->>T: User Data
T->>P: Fetch Trade History
P-->>T: Trade History
T->>L: Classify strategy (history)
L-->>T: strategy_analysis (JSON)
T-->>S: strategy_result
S-->>C: Tool Response (JSON)find_top_tradersFetch top-performing traders from the Polymarket leaderboard with bot detection.
Parameter | Type | Description |
| integer | Number of traders (1ā50) |
| string |
|
Output: Array<{ profile_id, pnl, is_bot }>
analyze_trader_strategyDeep-dive analysis of a single trader using trade history + LLM classification.
Parameter | Type | Description |
| string | Wallet address ( |
Output: { strategy_description, risk_level, risk_justification, success_score, is_bot }
generate_batch_reportConcurrent analysis of multiple profiles with error-resilient execution.
Parameter | Type | Description |
| string[] | Array of profile IDs (1ā50) |
Output: Array<{ profile_id, pnl, strategy_description, risk_level, risk_justification, success_score, is_bot }>
Node.js ā„ 22
npm ā„ 10
OpenAI API key (for strategy analysis)
# Clone the repository
git clone <your-repo-url>
cd polymarket-mcp-bot-analyst
# Install dependencies
npm install
# Configure environment
cp .env.example .env
# Edit .env and add your OPENAI_API_KEY# Build TypeScript
npm run build
# Start the MCP server (stdio transport)
npm start
# Or run directly with tsx (development)
npm run devAdd this server to your Claude Desktop configuration:
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"polymarket-bot-analyst": {
"command": "node",
"args": ["/absolute/path/to/polymarket-mcp-bot-analyst/dist/index.js"],
"env": {
"OPENAI_API_KEY": "sk-..."
}
}
}
}Edit %APPDATA%\Claude\claude_desktop_config.json:
{
"mcpServers": {
"polymarket-bot-analyst": {
"command": "node",
"args": ["C:\\path\\to\\polymarket-mcp-bot-analyst\\dist\\index.js"],
"env": {
"OPENAI_API_KEY": "sk-..."
}
}
}
}After saving, restart Claude Desktop. The three tools will appear in the tools menu (šØ icon).
The test runner executes all three tools against the live Polymarket API and generates the required JSON artifacts:
npm run test:runThis produces:
File | Description |
| Full execution log with data for 3+ traders |
| Latency metrics for each endpoint |
| Architectural description of each endpoint |
polymarket-mcp-bot-analyst/
āāā src/
ā āāā index.ts # MCP server entry point
ā āāā types.ts # Shared interfaces & config
ā āāā api/
ā ā āāā polymarket.ts # Polymarket Data API wrapper
ā āāā tools/
ā ā āāā traders.ts # find_top_traders handler
ā ā āāā analysis.ts # analyze_trader_strategy handler
ā ā āāā reports.ts # generate_batch_report handler
ā āāā utils/
ā ā āāā llm.ts # OpenAI LLM integration
ā āāā test-run.ts # Artifact generator script
āāā test_run.json # Generated test run log
āāā performance_report.json # Generated latency metrics
āāā my_report.json # Generated architecture report
āāā package.json
āāā tsconfig.json
āāā .env.example
āāā .gitignoreEnvironment Variable | Required | Description |
| Yes | OpenAI API key for GPT-4o-mini |
src/types.ts)Constant | Default | Description |
|
| API base URL |
|
| HTTP request timeout |
|
| Max retry attempts per request |
|
| Base delay for exponential backoff |
|
| Min trades to flag as bot |
|
| Min trades/hour for bot flag |
MIT
Analyze a trader's strategy using trade history and LLM classification. Returns strategy type, risk level, success score, and bot detection.
| Name | Required | Description | Default |
|---|---|---|---|
| profile_id | Yes | Polymarket profile ID ā wallet address (0xā¦) or username (@name). |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the burden of disclosing behavior. It mentions the methodology (trade history and LLM classification) and the return fields, which adds transparency. However, it does not disclose potential limitations (e.g., data freshness, latency, reliance on profile existence) or safety characteristics beyond the fact that it is an analysis function.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, well-structured sentence that front-loads the core action and immediately specifies what the tool returns. Every word earns its place; there is no redundant phrasing.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's low complexity (one parameter) and the absence of an output schema, the description provides adequate context by listing the return categories (strategy type, risk level, success score, bot detection). It does not detail output structures or error scenarios, but for a simple analyzer this is reasonably complete.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already provides a thorough description of the single parameter (profile_id) including format and examples. The description adds no additional parameter semantics, so the baseline score of 3 for high schema coverage applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's function with a specific verb ('Analyze') and resource ('a trader's strategy'), and distinguishes it from siblings by focusing on individual trader analysis rather than discovery (find_top_traders) or batch reporting (generate_batch_report). It also enumerates the key outputs, leaving no ambiguity about the tool's scope.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies use when a specific trader's strategy needs evaluation, but it does not explicitly state when to prefer this tool over alternatives or when not to use it. There is no mention of exclusions or comparisons with sibling tools, so the guidance is only implicit.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Fetch top traders from the Polymarket leaderboard. Detects bots based on trade frequency and volume.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | Yes | Number of traders to return (1ā50). | |
| timeframe | Yes | Leaderboard timeframe: "7d", "30d", or "all_time". |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the burden. It discloses the bot-detection behavior based on trade frequency and volume, which is useful. However, it does not detail return format, whether bots are filtered or flagged, or other operational aspects.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences long with no unnecessary words. The primary action is front-loaded, and the second sentence adds valuable behavioral context about bot detection.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple tool with two well-documented parameters and no output schema, the description is reasonably complete. It explains the main purpose and a key behavior, though it could clarify what data is returned and how bot detection affects the output.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, with both parameters clearly documented in the schema. The description does not add additional parameter-specific meaning beyond what the schema already provides, so the baseline 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool fetches top traders from the Polymarket leaderboard, with a specific verb and resource. It adds a unique capability (bot detection) that distinguishes it from siblings like analyze_trader_strategy and generate_batch_report.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The context is clear: use this to obtain leaderboard data. It implies a straightforward use case without explicit alternatives or exclusions, but the sibling tool names are distinct enough that the intended usage is evident.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Concurrently analyze multiple trader profiles and generate a combined report with PnL, risk, score, and bot status.
| Name | Required | Description | Default |
|---|---|---|---|
| profile_ids | Yes | Array of profile IDs (wallet addresses or usernames). |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It adds useful context by noting that analysis happens 'concurrently' and that the result is a 'combined report' with specific metrics. However, it does not explicitly state whether the operation is read-only, the nature of the report structure, or any potential side effects or limitations (e.g., rate limits, failure handling).
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single sentence that is front-loaded with the action ('Concurrently analyze') and resource ('multiple trader profiles'), followed by the output contents. Every word contributes, with no redundancy or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the absence of an output schema, the description should clarify what the 'combined report' looks like (e.g., per-profile breakdown vs. aggregated summary) and any edge cases. It lists key fields but leaves structural ambiguity about how results are organized, making it minimally complete but not fully self-sufficient.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% for the single parameter profile_ids, with a clear description in the schema. The tool description merely echoes 'multiple trader profiles,' adding little beyond the schema's 'Array of profile IDs (wallet addresses or usernames).' Thus the baseline of 3 is appropriate.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's function: 'Concurrently analyze multiple trader profiles and generate a combined report' with specific output components (PnL, risk, score, bot status). This distinguishes it from sibling tools like analyze_trader_strategy (single profile) and find_top_traders (discovery).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The phrase 'multiple trader profiles' implies a batch use case, but it does not explicitly contrast with analyzing profiles individually via analyze_trader_strategy, nor does it state conditions for when this tool is preferred. Usage context is only implied, not explicitly guided.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v1.0.0analyze_trader_strategyfind_top_tradersgenerate_batch_reportScored across 3 tools
Each tool has a clearly distinct purpose: finding top traders, analyzing a single trader's strategy, and generating batch reports. No overlap in functionality.
All tool names follow the verb_noun pattern (find_top_traders, analyze_trader_strategy, generate_batch_report), providing a predictable and consistent naming convention.
Three tools are perfectly scoped for this niche domain of trader analysis and bot detection. Each tool earns its place, and the count is neither too thin nor excessive.
The tool set covers the full workflow: discovering traders via the leaderboard, deep-diving into individual strategies, and scaling to batch analysis. No critical gaps or dead ends.
Research-only MCP server: your AI as a quant research desk. 90 tools, no trades, no brokers.
Read-only MCP server for live Polymarket, Kalshi, Limitless odds; Manifold sentiment.
MCP server for Gainium ā manage trading bots, deals, and balances via AI assistants
Crypto market intelligence, token rug-checks, and wallet verification in one MCP server.
