
varrd
by augiemazza·★ 17·Score 43
AI-powered trading research platform with MCP server providing 9 tools for market edge validation.
Overview
VARRD is an AI-powered trading research platform that turns trading ideas into quantitative formulas using a domain-specific language. It provides MCP server with 9 tools for browsing the edge library, conducting AI research, autonomous discovery, and managing credits. The platform statistically validates market behaviors across stocks, futures, and crypto using proper guardrails to prevent overfitting and p-hacking. The MCP integration allows users to ask questions like 'What VARRD edges are firing right now?' or 'What happens to gold when silver ETFs are making 100-day new lows?' directly to their AI assistants.
Try asking AI
After installing, here are 5 things you can ask your AI assistant:
When to choose this
Choose VARRD if you're a trader or quantitative researcher who needs statistically validated market analysis with transparent methodology and protection against common backtesting pitfalls.
When NOT to choose this
Avoid VARRD if you need unlimited historical data access, lower-tier pricing for casual trading research, or open-source models without vendor lock-in.
Tools this server exposes
9 tools extracted from the READMEvarrd_edgesBrowse the validated edge library with filters
varrd_aiMulti-turn research to test any trading idea
autonomous_varrd_aiAutonomous discovery - VARRD finds edges for you
searchSearch your saved strategies
get_hypothesisGet full details on your own strategy
check_balanceCheck credit balance and detect completed payments
buy_creditsPurchase credits via card or USDC on Base
get_briefedGet personalized market news tied to your edge library
reset_sessionKill a stuck research session
Comparable tools
Installation
Install via pip: pip install varrd
Configure in Claude Desktop:
{
"mcpServers": {
"varrd": { "url": "https://app.varrd.com/mcp" }
}
}FAQ
- How does VARRD prevent overfitting?
- VARRD uses out-of-sample testing that's permanently locked after one test, K-tracking to count every test variant, Bonferroni correction for multiple comparisons, and vectorized embeddings to detect duplicate hypotheses.
- Can I test my own trading ideas?
- Yes, VARRD provides a research tool where you can input any trading idea, which will be converted into a quantitative formula and tested statistically with proper guardrails.
On Hacker News
Recent discussion from the developer community.
- Story by varrd1 · 2026-03-08
Compare varrd with
Last updated · Auto-generated from public README + GitHub signals.