BTCBTC$64,892-0.05%|
ETHETH$1,825.53+0.16%|
USDTUSDT$0.99937+0.01%|
BNBBNB$592.7000+0.19%|
USDCUSDC$0.99966+0.00%|
XRPXRP$0.96160-0.02%|
SOLSOL$72.5900+1.15%|
TRXTRX$0.30499+0.46%|
FIGR_HELOCFIGR_HELOC$1.0390+0.00%|
HYPEHYPE$69.9800+5.40%|
ZECZEC$914.5500+2.68%|
XMRXMR$496.1200+1.16%|
DOGEDOGE$0.05905+0.27%|
LEOLEO$9.3300+1.33%|
USDSUSDS$0.99996+0.12%|
WBTWBT$62.7900-0.21%|
LINKLINK$9.1700-0.16%|
ADAADA$0.15398-0.02%|
XLMXLM$0.13874-0.96%|
USDEUSDE$0.99940-0.01%|
DAIDAI$0.99996-0.01%|
RAINRAIN$0.00619833+80.63%|
NEARNEAR$3.2400+1.76%|
BCHBCH$208.3100-0.30%|
LTCLTC$44.7900-0.05%|
CCCC$0.08225-1.18%|
AVAXAVAX$7.1700-0.20%|
UNIUNI$5.2300+0.98%|
XAUTXAUT$3,848.34-0.23%|
SUISUI$0.72453-0.31%|
BTCBTC$64,892-0.05%|
ETHETH$1,825.53+0.16%|
USDTUSDT$0.99937+0.01%|
BNBBNB$592.7000+0.19%|
USDCUSDC$0.99966+0.00%|
XRPXRP$0.96160-0.02%|
SOLSOL$72.5900+1.15%|
TRXTRX$0.30499+0.46%|
FIGR_HELOCFIGR_HELOC$1.0390+0.00%|
HYPEHYPE$69.9800+5.40%|
ZECZEC$914.5500+2.68%|
XMRXMR$496.1200+1.16%|
DOGEDOGE$0.05905+0.27%|
LEOLEO$9.3300+1.33%|
USDSUSDS$0.99996+0.12%|
WBTWBT$62.7900-0.21%|
LINKLINK$9.1700-0.16%|
ADAADA$0.15398-0.02%|
XLMXLM$0.13874-0.96%|
USDEUSDE$0.99940-0.01%|
DAIDAI$0.99996-0.01%|
RAINRAIN$0.00619833+80.63%|
NEARNEAR$3.2400+1.76%|
BCHBCH$208.3100-0.30%|
LTCLTC$44.7900-0.05%|
CCCC$0.08225-1.18%|
AVAXAVAX$7.1700-0.20%|
UNIUNI$5.2300+0.98%|
XAUTXAUT$3,848.34-0.23%|
SUISUI$0.72453-0.31%|
Adjacent (adjacent.markets)
Tool / APITier-1Index provider and data APIVerified

Adjacent (adjacent.markets)

The independent index and data house for event contracts: rules-based prediction market indices, reference rates, market data, news and CFTC filings in one API and MCP server.

Visit website

Key Facts

CategoryIndex provider and data API
AccessFree tier: public /api/v1/public/* endpoints need no token and return data delayed by 15 minutes, with limited site data and MCP access. Pro is $50/mo (5 seats, 30 requests/minute, 20k requests/day) and Premium $250/mo (10 seats, 180 requests/minute, 150k requests/day) for realtime endpoints, full history and authenticated MCP; an Adjacent Press subscription is included on paid plans. Request limits are organisation-wide — every API key and signed-in session shares the same budget, MCP included. Realtime calls use a bearer API key created under Settings → API keys, and a key can only reach the endpoints its scopes cover.
Venues CoveredKalshi and Polymarket markets, prices, trades, quotes and candles — plus indices, reference rates, news and CFTC filings built from them
Data ConfidenceVerified
Last Reviewed30 Sep 2026

What It Does

An independent index provider and data company for event contracts — the neutral calculating agent behind rules-based prediction market benchmarks rather than a venue or a dashboard. It exists because the asset class is fragmented: the same event can be listed ten different ways across ten platforms, with inconsistent settlement rules and no standardisation, so institutional capital wanting to hedge event risk and sophisticated retail wanting direct event exposure have had no passive, comparable instrument to buy. Adjacent builds that layer — thematic, narrative, categorical and return-based indices, reference rates that put the same market across venues or rolling contracts on a single price series, and the underlying market data — over Kalshi and Polymarket, with governance documents and a published methodology for every ticker. Levels are readable in a defined way: probability-based political indices use a 50–150 scale where 100 is neutral and the level minus 50 gives the implied win probability, and total-return indices start at 100 and measure cumulative return instead. Data reaches users through the site, a public JSON API, an MCP server for AI agents, embeddable widgets and CSV settlement exports, and the news arm that began life as Adjacent News is now folded into the same subscription as Adjacent Press.

Overview

What this is

Adjacent (adjacent.markets) is an independent index provider and data company for prediction markets and event contracts — a calculating agent that publishes rules-based benchmarks over Kalshi and Polymarket markets, rather than a venue or a dashboard. It does not take orders, hold funds or list contracts. What it produces is the layer the asset class has been missing: indices, reference rates and the market data underneath them, plus a news desk and a CFTC filing tracker built from the same coverage.

The same data is delivered through the site, a public JSON API, an MCP server for AI agents, embeddable widgets and CSV settlement exports.

The problem it exists to solve

Prediction markets are fragmented. The same event can be listed ten different ways across ten platforms, with inconsistent settlement rules, and there is no standardisation between them. That leaves two groups unserved at once: institutional capital that wants to hedge event-driven risk, and sophisticated retail that wants direct event exposure without choosing a platform and a contract, deciding on strikes and expiries, and then tracking resolution and settlement themselves. A rules-based index is the standard answer to that problem — one number, one method, published.

What it publishes

  • Indices — thematic, narrative, categorical and return-based, at both composite and single-race level. The political family is the clearest example: the red and blue composites, their component sub-indices (presidential, Senate, House, governor, mayoral), total-return variants such as red_tr, and NFL team indices published under the nti methodology.
  • Reference rates — one price series for the same underlying where it appears on more than one venue, or where it rolls from one contract into the next.
  • Market data — markets and events, with prices, trades, quotes and candles.
  • News — stored articles connected to the markets, indices and filings they relate to. The news desk began life as Adjacent News and is now folded into the same subscription as Adjacent Press.
  • CFTC filings — industry filings with status changes, attachments and related markets.

Every ticker carries a governance document and a published methodology, which is the point: an index is only useful if you can read why its level is what it is.

Reading the numbers

The scales are not interchangeable, and that is where most misreadings happen.

  • Political indices use a 50–150 scale, not 0–100: value = 50 + the party's aggregate win probability in percent. So 50 is 0%, 100 is an even 50/50 and 150 is 100%. Subtract 50 from the level to get the implied probability — 95 is roughly 45%, 109 is roughly 59% and 130 is roughly 80%. A red level above 100 means Republicans are favoured across its races; above 100 on blue means Democrats are.
  • Total-return indices start at 100 and measure cumulative return instead — 125 on red_tr is a cumulative return of 25%, and the series is deliberately not smoothed.
  • NFL team indices start at 1000 and compound net asset value rather than expressing a probability at all.
  • A constituent market's price is a different scale entirely — 0–100% Yes probability — so a market price should never be compared directly with an index level.

Who it is for

  • Institutional and desk research treating event-contract exposure as an asset class rather than a one-off bet.
  • Sophisticated retail who want a benchmark rather than a single contract.
  • Builders and quants who need one normalised feed across Kalshi and Polymarket instead of reconciling two venues themselves.
  • AI agents that need reliable market data through MCP rather than scraped pages.
  • Newsrooms and analysts who need a citable level with a documented method behind it.

What it is not

It is not an exchange, a broker or a trading venue, and nothing here is investment advice. The v1 plans also do not grant redistribution rights or permit use in a financial product — that is a licensing conversation rather than a plan upgrade. Bitnxt is not affiliated with Adjacent.

How It Works

Start with public data, then switch to realtime

Every endpoint lives under https://api.adjacent.markets/api/v1/. The anonymous /api/v1/public/* tier needs no token and returns data delayed by 15 minutes — enough to explore the shape of the catalogue and build a prototype. Realtime endpoints need a bearer token: either a session token from the app or an API key created under Settings → API keys, sent as an Authorization: Bearer ak_... header or as an api_key query parameter. Keys are scoped, so a key can only reach the endpoints its scopes cover and anything outside them returns 403.

Choose the surface that matches the question

SurfaceUse it forAccess
Market dataFinding markets and events, then reading prices, trades, quotes and candlesPublic (delayed) or realtime with a key
IndicesA single rules-based measure built from many markets, with constituents and weightsRealtime with a key
Reference ratesOne series for the same underlying across venues or across contract rolls, with the spread between sourcesRealtime with a key
NewsStored articles linked to the markets, indices and filings they coverRequires an API key
FilingsCFTC industry filings with status changes, attachments and related marketsRealtime with a key
MCP serverLetting an AI agent search and read the same data without learning the REST APIPublic delayed, or realtime with a key on the URL

How an index is built

A composite is a fixed-weight blend of sub-indices, and the sub-indices are built from their own constituent markets. The current political catalog defines two composites:

CompositeTickerSub-indices and fixed weights
RepublicansREDPresidential PRESR 40%, Senate SENR 30%, House HOUSER 20%, Governor GOVR 8%, Mayoral MAYORR 2%
DemocratsBLUEPresidential PRESD 40%, Senate SEND 30%, House HOUSED 20%, Governor GOVD 8%, Mayoral MAYORD 2%

Weights inside each sub-index follow the UPFI methodology: presidential and Senate weights use lifetime cumulative contract or share count and time-to-resolution, while House, governor and mayoral weights use jurisdiction population — state population for the first two, city population for mayoral races — with no impact factor. RED and BLUE publish a trailing one-hour simple moving average of the raw composite value, while sub-indices publish their raw per-cycle value. That single detail matters before comparing a composite with its components: the composite is smoothed and the components are not.

Reading an index over time

A detail request returns the current level with 1-day and 7-day changes. Constituents come with a kind — market for leaf markets, index for a composite's child sub-indices — alongside the computed weight, which is how you explain what actually moved the level. Price history takes an interval and page size, for example /api/v1/indices/red/prices?interval=1hour&per_page=24, and the same series can be exported to CSV or TSV for offline work through the /export/ routes. Family-specific routes carry the official daily series and membership events such as admissions and settlements; for the NFL family the index detail also returns holdings as sleeves plus the latest official daily print, while other families return an empty list.

Total-return and NFL indices break the probability convention

red_tr (RED-TR) is a path-dependent total-return index over RED's eligible constituents: base value 100, cumulative return measured from there, and deliberately not smoothed. NFL team indices (nti) work differently again. Each leaf index tracks one team across four sleeves — Super Bowl, conference, division and win total — and at each reconstitution every sleeve is sized to an equal dollar share of NAV, with the wrapping win-total pair splitting 125 each. The level starts at 1000 and compounds NAV period to period, floored at zero. A composite of team indices is an equal-weight blend of the children's raw NAV, again not a probability and not smoothed. The practical rule: read the methodology before comparing any two index levels, because three families on this platform use three different conventions.

Reference rates, spreads and rolls

A reference rate answers a different question from an index: not what a basket is worth, but what one underlying is worth across venues, or across the contracts it rolls through. The rate is the combined value of its sources and it reports the spread between them — a small spread means the sources agree, a large one means inspecting the individual markets before treating the rate as a single signal. The source list shows each market's contribution and weight.

  • volume_weighted_average — each source's effective weight is its configured or roll weight multiplied by its cumulative contract or share count: Σ(price × weight × count) / Σ(weight × count).
  • Claim-price and sleeve rates — event-contract and NTI sleeve rates weight on volume alone: Σ(price × count) / Σ(count), and a source that did not enter the print keeps its catalog weight while omitting its latest price.
  • simple_average — every source counts equally.
  • median — the middle source price, less sensitive to an outlier.

Chains are what keep a benchmark alive across expiries: successive contracts, such as the 2024 and 2028 presidential contracts, are joined, and a roll schedule blends adjacent contracts during the roll window. Passing sources=true with price history returns the per-leg roll weight at each point, so the seam in the series is visible rather than implied.

Using it from an AI agent

The MCP server at https://mcp.adjacent.markets/mcp exposes four tools rather than the whole REST surface. The public URL serves the 15-minute-delayed tier; appending an API key as apiKey on the URL switches it to realtime.

ToolWhen to use it
listBrowsing tradable events (the default), markets, indices, rates, filings, announcements or news. Search is all-words AND, and news needs a key. Historical catalog browsing passes include_resolved with created_before or as_of.
findDiscovering an entity by topic when you have no id, optionally constrained by type — index, rate, event, market, news, filing, announcement or similar. type=similar accepts a platform to rank only that venue.
getFull detail by id; indices include their constituents. Index, rate and market reads accept as_of for a historical snapshot.
pricePrice history — a human timeframe live (24h, 7d, 30d, 90d), or end / as_of for a past question. A live timeframe is not a substitute for end when the question is about the past.

Ids chain from one tool to the next: markets use <platform>:<raw>, for example kalshi:KXPRESPARTY-2028-R, while indices and rates use slugs such as red. Two limits are worth designing around: the API does not store historical membership, so historical event reads are refused, and failures carry a stable code — RATE_LIMITED, FORBIDDEN, NOT_FOUND, SEARCH_NO_MATCH, NO_PRICE_AT_AS_OF or CREATED_AFTER_AS_OF — so an agent can branch on the reason instead of retrying blindly.

Method Notes & Limitations

Where the data comes from

Market data is drawn from Kalshi and Polymarket — markets and events, prices, trades, quotes and candles — and every derived product is computed from those prices: indices from rule-defined constituent sets and weights, reference rates from the same markets across venues or across rolls, news from stored articles linked to the markets, indices and filings they cover, and filings from CFTC industry documents with their status changes and attachments. Methodology and governance documents are published per ticker and per family, and the documentation site includes an agent-readable index at /llms.txt, a single-file version at /llms-full.txt, a downloadable /docs.zip and a changelog.

How the derived values are computed

  • Index composites are fixed-weight blends of sub-indices (RED and BLUE at 40 / 30 / 20 / 8 / 2), and each sub-index is built from constituent markets weighted by the UPFI methodology — contract or share count and time-to-resolution for presidential and Senate races, jurisdiction population for House, governor and mayoral races, with no impact factor.
  • Smoothing differs by level: RED and BLUE publish a trailing one-hour SMA of the raw composite, sub-indices publish raw per-cycle values, and the total-return and NFL families are not smoothed at all.
  • Reference rates are volume- or weight-weighted combinations of their sources, and the reported spread is the distance between those sources. Different methods produce different values from the same inputs, so the method is part of reading the number.
  • Scales are per family: 50–150 probability for the political indices, 100-based cumulative return for the total-return series, 1000-based NAV compounding for NFL team indices, and 0–100% for an individual market price.

Known scope limits

  • The public tier is delayed by 15 minutes and needs no token; realtime requires a key.
  • Free-tier price endpoints serve hourly history for the previous 30 days or daily history for the previous 90 days. Finer intervals and deeper history are paid.
  • Paid request limits are organisation-wide and reset at 00:00 UTC — every API key and signed-in session draws on the same budget.
  • WebSockets, the index builder and ratings data are not included in the v1 plans.
  • News through the API and MCP requires a key, and because the API does not store historical membership, historical event reads are refused.
  • Index, rate and market historical snapshots are available with as_of, which is the supported way to answer a question about a past point in time.

Cost and limits if you build on it

PlanPriceDataSeatsLimits
Free$0Public API and MCP, 15-minute delayed—50 requests / 10 seconds / IP at the edge
Pro$50/monthRealtime indices, rates, markets, history, research and MCP530 requests/minute; 20,000/day
Premium$250/monthEverything in Pro with higher throughput10180 requests/minute; 150,000/day

Adjacent Press is included on paid plans for seat emails, and free at the content level. Plans, with live usage, are managed under Settings → Billing & usage.

Licensing

Neither Pro nor Premium permits redistribution of the data or use in a financial product — the published plans are for research and internal automation. Data licensing is a separate conversation with Adjacent, so if you intend to publish derived figures, check the terms and the per-ticker governance document first.

Housekeeping worth knowing

The product line has moved around: the news desk that began as Adjacent News is now part of the Adjacent Press subscription, the old adj.news domain redirects to the current site, and the legacy documentation domain docs.adj.news no longer resolves — current documentation lives at docs.adjacent.markets. Older links and third-party write-ups may therefore point at names that no longer exist.

Editorial note

Bitnxt is not affiliated with Adjacent. Levels, formulas and plan limits on this page were read from Adjacent's own site and documentation on the review date and change as the platform moves, so verify any figure against the live documentation before relying on it. This page carries no trading links, no exchange referrals and no investment advice.

Venues Covered

Kalshi and Polymarket markets, prices, trades, quotes and candles — plus indices, reference rates, news and CFTC filings built from them

Access & Pricing

Free tier: public /api/v1/public/* endpoints need no token and return data delayed by 15 minutes, with limited site data and MCP access. Pro is $50/mo (5 seats, 30 requests/minute, 20k requests/day) and Premium $250/mo (10 seats, 180 requests/minute, 150k requests/day) for realtime endpoints, full history and authenticated MCP; an Adjacent Press subscription is included on paid plans. Request limits are organisation-wide — every API key and signed-in session shares the same budget, MCP included. Realtime calls use a bearer API key created under Settings → API keys, and a key can only reach the endpoints its scopes cover.

Frequently Asked Questions

Sources & Verification

The factual claims on this page were checked against the sources below.

Bitnxt uses cookies to run the site, understand how it is used, and measure ads. You can accept or decline below, and change your choice any time in our Privacy Policy.