
pdata MCP server
Which servers in the official MCP registry are pdata's own, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.
read 2026-10-21
- Registry nameworld.pdata/mcpverified in the MCP registry
The registry name sits under pdata.world. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-10-21
- Tools14
- Server calls itselfpdata — prediction markets
- Server version2.4.3
- Protocol version2025-06-18
On 2026-10-21 the server answered initialize and tools/list without a sign-in. First answer in 267 ms.
- Whether its tools do the product's job has not been judged yet.
- Checked on 6 of the 7 days it takes to judge reliability.
Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.
- INITE MCP scoreafter a weeka score needs 7 daily checks: 6 so far
- Answeredmeasuring6 of the 7 daily checks it takes to print a share
- Median first answer598 ms
- Space in an agent's context≈ 11,397 tokensthe definitions of all its tools (14), which an agent carries on every turn
- Tool definitionsA · 100 of 100
- Typed results0 of 14 toolsdeclared outputSchema: results come typed, read once
- Every tool passes all six checks.
What an agent can do with it
- not sure · compare odds for the same event across platforms
- not sure · track 24-hour trading volume on prediction markets
- not sure · pull prediction market data through an API
Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?
Worth a look
None of the five warning signs in these definitions.
A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (30%), doing the product's job (25%), definitions (20%), context (10%) and use (15%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →
list_eventsUse when an agent wants to browse, filter, or search prediction-market events across all 8 platforms. Returns paginated events with their child markets, prices, and platform metadata. Filterable by source…
get_eventUse when an agent needs the full detail of a single prediction-market event by its composite key (source + event_id). Returns the event with all of its child markets, latest probabilities, and metadata. The canonical URL…
get_event_similarUse when an agent needs events similar to a given event — cross-platform equivalents and topically-related markets, precomputed nightly from title embeddings. Best-first; empty list (never an error) when no similarities are known yet.
list_marketsUse when the user wants to browse, filter, or sort individual markets (not events). Returns paginated markets with their probability, volume, source, parent event ID, and 24h delta. Filterable by source, country, category, status, and…
get_marketUse when an agent needs the full detail of a single market by its composite key (source + market_id). Returns the market with its parent event reference, current probability, volume, 24h delta, and historical chart points. Includes a…
top_moversUse when the user wants to see markets with the largest 24h probability swings. Returns the top N movers (default 5, max 20) ranked by absolute `prob_24h_change`, among open markets that moved more than 2 points on over 1,000 units of 24h…
closing_soonUse when the user wants to see markets that will resolve within the next 36 hours. Returns markets sorted by close timestamp ascending. Takes the /markets filters, e.g. `source=` and `categories=` (CSV); price-derivative markets are…
search_suggestUse when an agent has a name or phrase and needs the `(source, id)` of the matching event or market before calling another tool. Resolves free text against events AND markets in one call, best-first, and reports `total_events` /…
get_summaryUse when an agent needs platform-level context before drilling into individual markets: one row per platform (Polymarket, Kalshi, Manifold, Myriad, Limitless, Predict, Opinion, Gemini) carrying 24h notional volume, active event and market…
list_countriesUse when an agent needs the list of country tags that markets have been tagged with. Returns the canonical country vocabulary with per-country market counts. Accepts the same filter params as `list_markets` (source, category, status…
list_news_moversUse when the user wants to see the news-feed view: markets that exhibited a probability spike in the chosen lookback window, scored by `|Δprob_pp| × log10(1+USD-equivalent volume)`. Refreshed after each Polymarket ingest cycle, typically…
news_movers_by_categoryUse when the user wants the news feed bucketed by category. Returns the same data shape as `list_news_movers` but pre-grouped into the canonical 14-category vocabulary, with cluster grouping inside each bucket. Useful for 'what's moving…
map_heatUse when you need a geographic snapshot of where prediction-market attention is concentrated. Returns per-country heat (capped news score, summed) for the current moment (`at=now`) or a past hour (`at=<ISO8601>`).
map_timelineUse when you need the hourly total-heat time series that drives the /map playback scrubber. History accretes from launch (no backfill).
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"pdata": {
"type": "http",
"url": "https://api.pdata.world/mcp"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http pdata https://api.pdata.world/mcpclaude.ai, ChatGPT and other clients: add https://api.pdata.world/mcp as a custom connector in the settings.
OpenAPI spec
21 paths, read 2026-10-21. Open the spec ↗
Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.
Cookies
Before we count this visit
Analytics and ad measurement tell us which pages and which of our ads do their job: Google Analytics and Google Ads, the Meta pixel, Yandex Metrica, and one first-party id of our own. Nothing is stored until you answer, and declining changes nothing about the site.