About voxodds
VoxOdds is a remote MCP server for agents and clients that need live prediction-market data from Polymarket and Kalshi. It combines executable quotes, fee-aware EV checks, reviewed market pairs, and scored forecasting records.
Key features
- 01Live odds and trending markets
- 02Cross-platform market comparisons
- 03Depth- and fee-aware executable quotes
- 04EV checks and edge signals
- 05Forecast submissions and scorecards
Use cases
- 01Find live odds for a market question
- 02Compare equivalent Polymarket and Kalshi markets
- 03Check a proposed bet against market-based EV
- 04Track forecast accuracy against market prices
MCP support
Exposes prediction-market research and comparison tools to MCP clients over a stateless Streamable HTTP endpoint. The server also offers a REST/OpenAPI alternative.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://voxodds.com/mcpclaude mcp add --transport http com-voxodds-voxodds https://voxodds.com/mcp{"mcpServers":{"com-voxodds-voxodds":{"type":"http","url":"https://voxodds.com/mcp"}}}Open in Cursor{"servers":{"com-voxodds-voxodds":{"type":"http","url":"https://voxodds.com/mcp"}}}Open in VS Code{"mcpServers":{"com-voxodds-voxodds":{"type":"http","url":"https://voxodds.com/mcp"}}}{"mcpServers":{"com-voxodds-voxodds":{"serverUrl":"https://voxodds.com/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
18| Parameter | Type | Required | Description |
|---|---|---|---|
| outcome | string | Yes | Exact outcome name of that market, e.g. 'Yes' |
| market_id | string | Yes | Market id from list_trending_markets or get_market_odds |
| probability | number | Yes | Your probability for that outcome, 0.01-0.99 |
| forecaster_id | string | Yes | Your stable public handle: 3-32 chars, a-z, 0-9 and dashes |
Details
From the project READMEVoxOdds MCP server
Remote Model Context Protocol server for VoxOdds — live prediction-market odds (Polymarket, Kalshi) with all-in executable quotes, fee-aware EV checks, verified resolution rules, and an audited AI-vs-market forecasting track record. Free, no API key.
- Endpoint:
https://voxodds.com/mcp(Streamable HTTP, JSON-RPC 2.0, stateless, no auth) - Official registry:
com.voxodds/voxodds· Glama connector: https://glama.ai/mcp/connectors/com.voxodds/voxodds - REST alternative: OpenAPI at
https://voxodds.com/api/openapi.json, Swagger UI athttps://voxodds.com/docs, agent notes athttps://voxodds.com/llms.txt
Use it
Claude Code:
claude mcp add --transport http voxodds https://voxodds.com/mcp
mcp.json (Cursor, VS Code, and other clients that read this format):
{
"mcpServers": {
"voxodds": { "url": "https://voxodds.com/mcp" }
}
}
Claude.ai / ChatGPT — add https://voxodds.com/mcp under Settings → Connectors.
Try it with curl:
curl -s https://voxodds.com/mcp \
-H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_market_odds","arguments":{"question":"Will Lando Norris be the 2026 F1 champion?"}}}'
Tools
| Tool | What it does |
|---|---|
get_market_odds(question) | Resolve a natural-language question to one live market and return its implied probabilities. Refuses ambiguous matches. |
list_trending_markets(category?, search?, limit?) | Trending markets by volume, optionally filtered. |
compare_platforms(query?, limit?) | Same event priced on Polymarket vs Kalshi, with the spread. |
get_executable_quote(platform, market_id, side?, amount_usd?) | All-in fill for one exact contract: order-book depth, venue taker fees, average price. |
compare_executable_quotes(polymarket_id, kalshi_ticker, ...) | Executable comparison across venues for a reviewed equivalent pair. |
list_executable_opportunities(amount_usd?, side?) | Ranked all-in opportunities across YES and NO of every reviewed equivalent pair. |
list_arbitrage_candidates(max_capital_usd?) | Size-matched opposite-side coverage whose captured all-in cost is below the payout (unfilled candidates; legs are non-atomic). |
list_sportsbook_surebets(sport_key?, capital_usd?) | Sportsbook surebets; returns a disabled notice until a live odds feed is configured. |
check_bet(event, side, odds_offered?, sport?) | Pre-bet check: conservative event resolution, then EV, break-even odds and Kelly against the market reference. |
find_best_price(event, side, sport?) | Best available price across configured books; prediction-market midpoints are never presented as executable. |
get_edge_signals(limit?) | Current edge signals from the VoxOdds edge engine. |
get_research_theses() | Research-desk theses with supporting markets. |
get_track_record() | The audited AI-vs-market experiment: Brier scores vs contemporaneous prices, losses included. |
submit_forecast(market_id, outcome, probability, forecaster_id) | Record your own probability on a live market (append-only) and build a public Brier track record. |
get_forecaster_record(forecaster_id) | Public scorecard for any forecaster id. |
get_world_cup_odds() / get_world_cup_brief() / get_world_cup_matchday(date?) | Archived World Cup 2026 tools. |
Related
- Live comparison page: https://voxodds.com/polymarket-vs-kalshi
- Open daily dataset of the same price gaps (CSV, CC BY 4.0): https://github.com/softdevfz/polymarket-kalshi-price-gaps
- Bluesky feed of prediction-market posts: https://bsky.app/profile/voxodds.com/feed/prediction-markets
Notes
- Displayed probabilities are not executable quotes; verify order books, fees, fills, eligibility and resolution rules before acting. Research only; check venue eligibility and jurisdiction.
- Venue links returned by the tools may carry a referral code (the price is the same for you).
- Scored-forecast dataset:
https://voxodds.com/api/v1/scoreboard/dataset.csv(CC BY 4.0). - Operated by Soft Dev FZ LLC (Fujairah Creative City, UAE). Contact: [email protected]
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.