Discover tools for AI investment research

For crypto researchers who need to inspect a token’s on-chain trading, compare activity or follow wallet fund flows. Connect the hosted MCP and authorize a Bitquery account, then ask for token volume, OHLC candles, recent trades for a wallet or cross-chain fund paths; the service returns structured blockchain data. It focuses on transaction-level DEX and wallet activity, while CryptoQuant focuses more on on-chain metrics and research, so they serve different workflows. Official docs say the service is read-only and cannot trade or change blockchain state. Listed coverage includes Solana, Ethereum, BSC, Base, Arbitrum, Optimism, Polygon, Tron and Robinhood Chain. Most clients use OAuth with a free account; Codex desktop setup specifically requires a Bitquery bearer token. Official pricing lists a 7-day trial with 100 MCP credits, not a permanent free tier. Trading MCP provides 25,000 credits/month at a $13/month equivalent billed yearly ($19/month on monthly billing); AI Investigation MCP provides 200,000 credits/month at a $100/month equivalent billed yearly ($149/month monthly). The repository README confirms these are annual-billing rates; check the checkout page for current billing terms. Historical windows, archive access and the required plan must be confirmed against MCP entitlements; do not assume API-plan add-ons apply. Terms can change; this review did not connect or query the service.

Coding required: No coding required

For crypto investors who want on-chain evidence but do not want to hunt for API endpoints or interpret metric definitions manually. Connect the hosted MCP server and complete OAuth, then ask about exchange netflows, miner activity or network indicators. The agent can discover endpoints, read metric descriptions, query time series and search CryptoQuant Research or QuickTake. The official page lists BTC, ETH, XRP, TRX, stablecoins, ERC-20 tokens and other altcoins, with 200+ on-chain metrics across 490+ assets. Its added value is combining metric interpretation and data retrieval from the official provider. The Basic free plan lists 10,000 monthly credits, 10 requests/minute and 30-day history, but excludes on-chain data; higher tiers unlock more. Research content is free during the beta. The page labels its sample answers illustrative, not live. This review did not connect to or query the server.

Coding required: No coding required

The official repository provides stock, option, crypto and portfolio tools subject to account and API access. Documentation distinguishes local stdio from remote-deployment restrictions; this directory does not connect accounts or place orders.

Supports live execution

Official CLI for trading, accounts, and market data, with JSON output and paper/dry-run defaults

Supports live execution

Official Alpaca skills for backtesting, paper trading, and Broker API workflows

Supports live execution

Hosted tools for account, portfolio, risk, market data, and trade instructions with user review before submission

Tiantian Fund's official tutorial says users can search for “SKILL” in the Tiantian Fund app, copy setup content, and configure it in compatible tools such as OpenClaw, Claude Code, QClaw, and AutoClaw; the tutorial demonstrates WorkBuddy. Public materials mainly describe fund lookup and Q&A. We did not log into an account, install the Skill, or verify whether it is read-only or can trade. Account, API-key, and regional eligibility requirements may affect access.

GF Securities' official page and public frontend script confirm a personal access-token portal for E-Taojin Skills/MCP and show a flow that requests a Skills token after SMS login. The page does not state that a securities account or real-name brokerage account is required, and it provides no public pre-login tool list. We did not log in or request a token; current issuance availability, compatible AI clients, and trading capability were not tested. The E-Taojin app itself supports securities trading, but this does not establish order placement through its Skills/MCP.

Guosen Securities' official Skills Center lists macro, market, and financial-data queries, fund comparison, stock screening, and ETF screening. The page says users must register or log in to download Skills and shows a dedicated API key; its public script defines official ZIP download links. It states compatibility with OpenClaw. The page does not explicitly say the Skills are read-only or prohibit trading, and it does not describe order placement. We did not log in, download or install the ZIP files, or test trading. Securities-account or real-name requirements are not specified on the page.

Ant Digital Technologies' official page describes a financial MCP service for AI applications, with fund, investment-research, and advisory capabilities. It does not say whether individuals can apply directly, and does not publish detailed integration steps, pricing, securities-account connectivity, or order-placement capability. We did not apply for, integrate, or test the service; eligibility and callable functions remain unverified. This is an Ant Digital Technologies product, not an Ant Fortune retail fund Skill.

For investors or researchers who need a quick view of A-share short-term sentiment, themes and company events. Configure the hosted MCP endpoint and obtain a provider API key, then ask about a ticker's prices, fundamentals, announcements, Dragon-Tiger List, capital flow or opening auction, or request a date-based market review. The project documents 57 read-only query tools and review workflows that combine several data views; six additional tools alter only the user's own watchlist. It also lists SEC disclosure lookups. Its distinctive value over a generic quote connector is its A-share-specific workflow for limit-up ladders, auctions, themes, flows and official disclosures. The README lists 50 free calls/day, paused every day from 09:15 to 10:30 Beijing time. That includes the 09:15–09:25 opening call auction, so the free tier cannot serve that auction window. README annual plans start at CNY 1,000; verify current rates on the official pricing page. The repository code license does not make the hosted data service open source. This review checked documentation and vendor pages but did not call the service.

Coding required: No coding required

For investors who want an AI-assisted US-stock workflow without assembling market-data endpoints themselves. After installing uv and Git and connecting the MCP server, ask an agent to analyze a ticker, compare 2–5 stocks, review a holding, or scan a watchlist of up to 25 symbols. You can supply your cost basis and purchase date; responses include valuation, trend, risks, price provenance and data-quality warnings. Documented decision gates can flag reasons to wait, and the server also offers rule-based position-size estimates. The supported decision universe is USD-priced US common stocks and USD ADRs; ETFs, preferred shares and non-USD listings are outside new-buy decision support. The project uses Yahoo Finance through yfinance and says no market-data API key is required. Data may be delayed or incomplete. This review read the documentation but did not install or run the tool.

Coding required: No coding required

Public.com skill template for stocks, ETFs, options, and crypto

Multi-chain smart-contract venue for spot swaps and aggregation

Perpetuals DEX on an L1 with a full API and wallet-based access

Hosted MCP for portfolio research and orders in a dedicated Robinhood Agentic account

Solana aggregator and perps venue with an API and SDK

OpenClaw trading skill for sub-second Hyperliquid execution with hardcoded risk guardrails

OKX APIs covering exchange functions and a DEX aggregator

MCP service for creating projects, running backtests and optimizations, and deploying algorithms