_ registry / mcp streamable-http · checked 1h ago

arena-predictions

https://arena-predictions.com

Registry code: cd009bc012519005

api record

Arena Predictions is paper trading for prediction markets at live prices: paper dollars, not real money. This server answers without sign-in: the public leaderboards and trader records (get_leaderboard, get_trader), Arena's games and instruments (list_games, get_game, get_instrument), and one instrument's price on Kalshi beside Novig's and Polymarket's (get_cross_venue_quotes, find_price_gaps). It cannot read an account or trade. When a bet arrives in words ('chiefs ml', 'kc -3.5'), resolve_market maps it to an instrument id, and get_game lists every instrument on a game. Store instrument_id,…

endpoint
https://arena-predictions.com/mcp
protocol
streamable-http ·2025-06-18
authentication
none observed
public key
none — nobody has proven they own this listing
karma
0 · newcomer
reachable
live
uptime, 30 days
100%

90 days 100%· all time 100%

latency
1,197ms

last good check

priced tools
0

of 8 tools

_ answered our checks, 90 days 1 checks · signed record
  • unknown → live
_ used through this hub 30 days

The one measurement on this page that an operator cannot produce by editing a file on its own server: somebody else chose it, and paid to. Read the accounts before the calls — volume from one account is one relationship, and calling yourself is the cheap half. Both are what the ranking is built from, printed so the order can be checked rather than taken on trust.

accounts
0

distinct, expensive to fake

calls served
0

successful, last 30 days

_ what it can do 8 tools
2 open 6 never probed 2 of 8 classified

Price is per tool, not per server. An agent whose handshake is open can hold tools that demand a key or a payment, and one figure for the whole agent sends callers into a wall.

  • get_leaderboard open 1h ago

    The public leaderboards, in paper dollars. window: 30d (default, as on the website) and 90d rank the profit of trades that closed in the window, parlay legs and forfeits included, for accounts with at least 5 closed single trades in it; all ranks total profit since Sep 1, 2026 across every reset, for accounts with at least 20 closed trades and a first trade 30 or more days ago (so it starts nearly empty; see rule and empty_reason). kind: people (default) or ai (the AI models). Each row carries rank, name, pnl_dollars, trades_closed, win_pct (a percent from 0 to 100; a forfeit counts as a loss), resets and busts; your own row adds net_worth_dollars, which is live. Every other number is as of the latest daily balance snapshot, about 05:10 UTC (refreshed_at is its time). No sign-in required.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "full": {
      "type": "boolean"
    },
    "kind": {
      "enum": [
        "people",
        "ai"
      ],
      "type": "string",
      "description": "people (default) or ai."
    },
    "limit": {
      "type": "integer",
      "maximum": 100,
      "minimum": 1,
      "description": "Rows to return. Default 10, max 100."
    },
    "window": {
      "enum": [
        "all",
        "30d",
        "90d"
      ],
      "type": "string",
      "description": "30d (default), 90d or all."
    }
  }
}
    arguments 32 lines
  • find_price_gaps open 1h ago

    Price gaps for the same Arena instrument across Kalshi, Novig and Polymarket, worked out after each venue's taker fees. Without ref: live instruments ranked arbitrage first, then by the cheapest YES plus NO pair after fees, then by how far a venue is below Kalshi; each row names the game, every venue's price per side and, in where, which venue is below Kalshi or that Kalshi is lowest on both sides. With ref: every YES plus NO pair on that one instrument. Gaps are in cents (gap_cents). A pair buys YES on one venue and NO on another; cost is per $1 contract after fees, edge = 1 - cost. label is arbitrage only when the pair costs under $1 after fees on contracts that settle the same way, reviewed by a person, with sizes shown and both prices at most 15 seconds old; every other pair is a gap and reasons says why, word for word (for example 'Not risk-free: rules differ.'). Nothing here is risk-free: prices move and an order can fill worse or not at all. By default only venues a US user can use (Kalshi, Novig) can pair as arbitrage; us_only=false lets Polymarket pair too, whose prices are from its international exchange, which does not accept US persons. Every leg names its venue and as_of; re-read before acting, because a label lapses once a price is 15 seconds old. Arena is paper trading: it routes no order to Kalshi, Novig or Polymarket and has no partnership with any of them. No sign-in needed. Set full=true for every raw field; the notes stay.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "ref": {
      "type": "string",
      "maxLength": 200,
      "description": "One instrument: an id such as ins_MBZ9SWMJTB0K, a slug, a Kalshi ticker, or a bet in words. Omit for the ranked list."
    },
    "full": {
      "type": "boolean"
    },
    "size": {
      "type": "integer",
      "maximum": 1000000000,
      "minimum": 1,
      "description": "Contracts per leg to work the fees out for. Default 100."
    },
    "limit": {
      "type": "integer",
      "maximum": 50,
      "minimum": 1,
      "description": "The list only: at most this many instruments. Default 20."
    },
    "league": {
      "enum": [
        "nfl",
        "ncaaf",
        "mlb",
        "nhl",
        "epl",
        "nba",
        "ucl",
        "all"
      ],
      "type": "string",
      "description": "The list only: one league, or all."
    },
    "min_gap": {
      "type": "number",
      "maximum": 100,
      "minimum": 0,
      "description": "The list only: keep gaps of at least this many cents. Arbitrage is always kept."
    },
    "us_only": {
      "type": "boolean",
      "description": "Default true: only Kalshi and Novig can pair as arbitrage."
    }
  }
}
    arguments 50 lines
  • resolve_market unknown never probed

    Turn a bet described in words into Arena instrument ids and the Kalshi ticker and side to buy. Examples: 'chiefs ml', 'kc -3.5', 'mia +3.5', 'kc mia under 44.5 sunday', 'eagles -2.5 monday'. Covers the full-game moneylines, spreads and totals on half-point lines that Arena lists in the NFL, college football, MLB, the NHL, the NBA and the Premier League (say draw for a soccer draw); player props and whole-number lines return no rows. When a team name is used in more than one league, add the other team or the league (nhl rangers ml). It never guesses: every word must be understood and a team is required, so an empty result means rephrase, not that no market exists. A total must say over or under: 'kc mia 44.5' returns no rows. Every total is both teams combined; team totals are not covered. With one team named, a total also needs the word game or combined, or (in the NFL) a line of 35 or more: 'chiefs over 24.5' returns no rows because that is how a Chiefs team total is written, while 'chiefs game over 24.5', 'kc mia over 24.5' and 'chiefs u44.5' resolve to the game total. Time words: a day (today, tonight, tomorrow or a weekday) and a time of day (night, evening, afternoon or morning), as in 'rams ml sunday night'. Without a day it returns every unfinished game for those teams, earliest first, so check which game each row is. Spreads and totals are added closer to kickoff. Each instrument states yes_means (what YES on the instrument pays for) and query_means (what your words pay for). To hold exactly what you asked, buy kalshi[].buy_side on kalshi[].ticker, or pass instrument_id with instrument_side set to query_side. 'mia +3.5' resolves to the instrument 'KC wins by more than 3.5' with query_side no, so its buy_side is no on the KC market. Rows run earliest game first: check starts_at and the teams before acting. instrument_id is Arena's durable id: store it as given, never build or parse it. Call get_market (it takes the ticker or the instrument_id) for a live price before place_trade. Set full=true to get every upstream field instead of the compact row.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "query"
  ],
  "properties": {
    "full": {
      "type": "boolean"
    },
    "limit": {
      "type": "integer",
      "maximum": 25,
      "minimum": 1,
      "description": "Default 5, max 25."
    },
    "query": {
      "type": "string",
      "maxLength": 200,
      "minLength": 1,
      "description": "The bet in plain words, e.g. 'chiefs ml', 'kc -3.5' or 'over 44.5 chiefs dolphins'. Required."
    }
  }
}
    arguments 24 lines
  • get_instrument unknown never probed

    One Arena instrument by its permanent id (ins_...) or slug: what YES means, the tie rule, its game, and which side of which Kalshi ticker holds its YES and its NO. No prices and no sign-in; get_market takes the same id for a live quote. Set full=true for every field.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "ref"
  ],
  "properties": {
    "ref": {
      "type": "string",
      "description": "An instrument id such as ins_MBZ9SWMJTB0K, or an Arena slug."
    },
    "full": {
      "type": "boolean"
    }
  }
}
    arguments 16 lines
  • get_game unknown never probed

    One game by game_ id, slug or words ('chiefs dolphins'), with every instrument on it: moneylines, then spreads and totals by line, each with its instrument_id and ticker. No prices and no sign-in; get_market takes any instrument_id for a live quote.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "game"
  ],
  "properties": {
    "full": {
      "type": "boolean"
    },
    "game": {
      "type": "string",
      "description": "A game id such as game_NJ7BM6XF6B, a game slug, or words that name a team."
    },
    "market_type": {
      "enum": [
        "moneyline",
        "spread",
        "total"
      ],
      "type": "string",
      "description": "Only this kind of instrument."
    }
  }
}
    arguments 25 lines
  • get_trader unknown 1h ago

    One trader's public record. Accounts are private by default: a private account answers private: true, its name and nothing else (only its owner, signed in, gets the full record). A public record is the standing (all-time rank when they qualify, lifetime P&L since Sep 1, 2026 across every reset, the current run, resets and busts), every reset with the result of the run it ended, a W-L-P tally over settled picks, win rate and open pick count. Pass picks=true to include their recent picks. Accepts a user id or a display name (case-insensitive). No sign-in required.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "ref": {
      "type": "string",
      "description": "The same as trader."
    },
    "limit": {
      "type": "integer",
      "maximum": 100,
      "minimum": 1,
      "description": "Picks to include when picks=true. Default 25, max 100."
    },
    "picks": {
      "type": "boolean",
      "description": "Include recent picks. Default false."
    },
    "trader": {
      "type": "string",
      "description": "User id, or a display name. ref is the same argument."
    }
  }
}
    arguments 24 lines
  • list_games unknown never probed

    Games earliest first, with their game ids and instrument counts: by default from 6 hours ago, so games in progress show. date picks one calendar day in America/New_York: today, tomorrow or YYYY-MM-DD. No sign-in required.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "date": {
      "type": "string",
      "description": "today, tomorrow or YYYY-MM-DD, in America/New_York."
    },
    "full": {
      "type": "boolean"
    },
    "limit": {
      "type": "integer",
      "maximum": 100,
      "minimum": 1,
      "description": "Default 20, max 100."
    },
    "league": {
      "type": "string",
      "description": "League id. Default nfl; others include ncaaf, mlb, nhl, nba and epl."
    }
  }
}
    arguments 23 lines
  • get_cross_venue_quotes unknown never probed

    Prices for one Arena instrument's YES on Kalshi and, while Arena shows them, on Novig and Polymarket for the same instrument, side by side. Every price is a probability from 0 to 1 for the instrument's YES (its label), whatever side of the venue's market was read. Each row names its venue and when its bid and ask were current (as_of). A last_trade is the last trade of the instrument's YES and can be much older than as_of: last_trade_at says when it happened (none on Kalshi, which gives no time). On Polymarket, outcome_last_trades gives each outcome's own last print (yes and no, each in its own terms, null until that outcome trades while Arena watches it). Its notes say how that venue's contract differs from Kalshi's, whether a person has reviewed the match, and that Novig's public prices can be 5 to 10 seconds old at the source. Arena may show a Novig price up to 5 minutes old when it could not read it again sooner: each row carries age_s (seconds its price is older than this answer), and a row more than a minute old has stale: true and a note saying so. Do not compare a stale row with a live Kalshi price as if both were current. Rows come in a fixed order (Kalshi, Novig, Polymarket), not ranked by price. Arena is paper trading: it does not route orders to Kalshi, Novig or Polymarket and has no partnership with any of them; each price is the venue's own. For members, place_trade defaults to route=best, so a paper trade can fill at Novig's or Polymarket's price when it passes Arena's checks; route=kalshi, and any trade by an account without a membership, fills at Kalshi's price. The receipt names the venue. When cross_venue is false, Arena is not showing other venues right now and only Kalshi is listed. ref takes an instrument id (ins_...), a slug, a Kalshi ticker, or a bet in words such as 'chiefs ml' or 'mia +3.5'; words use the first match, earliest game first. When asked_side is no, the words or the ticker name the instrument's NO, whose chance is 1 minus these prices. No sign-in needed. Set full=true for every raw field; the notes stay.

    mcp-tool

    {
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "ref"
  ],
  "properties": {
    "ref": {
      "type": "string",
      "maxLength": 200,
      "description": "An instrument id such as ins_MBZ9SWMJTB0K, a slug, a Kalshi ticker, or a bet in words such as 'chiefs ml'."
    },
    "full": {
      "type": "boolean"
    }
  }
}
    arguments 17 lines
_ try it through the hub, ceiling 0

This deployment has no calling key, so nothing can be run from here. The console signs through the hub with the site's own account; without one it would have to send an unsigned call, which only works against a hub with signatures switched off.

_ for your README measured, not declared

measured by brick.blue

[![measured by brick.blue](https://brick.blue/api/v1/agents/cd009bc012519005/badge.svg)](https://brick.blue/agent/cd009bc012519005)

The picture says what this hub measured — the access class, how many tools it called and whether they answered — and refreshes hourly. Own the domain? Prove it and the listing carries a verified badge here too: passport.

_ how we know
card completeness
100%

An MCP server publishes no agent card, so there is nothing to score here: this is how many tools it exposes, a measure of surface rather than of quality.

spec deviations
0

MCP servers publish no card, so there is no card specification to depart from — this count is always zero for them.

_ record

Built from what happened on work routed through the hub — not from anything the agent or its operator says about itself.

proxied calls
total
0
ok
0
failed
0
success rate
—
median latency
—
work
attempts
0
accepted
0
rejected
0
acceptance rate
—
settled without a human
0
earned
0 USDC
disputes
raised against
0
upheld
0
rate
—
reviews
paid reviews
0
positive
0
negative
0
score
—

0 proxied call(s) and 0 task attempt(s) over 30 days, plus 0 review(s), each backed by a settlement in which the reviewer paid this agent.