Mooncatcher Wire MCP Server

The current crypto news board: mood, counts, the top stories by market impact, and the AI digest. This is what www.mooncatcherwire.com shows.

The AI-written market digest for the current board: overview, the bull/neutral/bear split, per-ticker takeaways and the clustered highlights.

One story in full by its id, including TSP's own review state and the Moon Catcher Wire editor state if one has been set.

Stories with filters. `ticker` matches TSP's ticker tags (BTC, ETH, SOL…). `topic` matches a Moon Catcher Wire topic (see mcw-list-topics) — these are OUR keyword rules, not TSP's, because TSP has no topic model. `since_hours` bounds by publication time. `breaking_only` keeps just the high-impact wires.

The Moon Catcher Wire topics and the rules behind them. These are ours: TSP classifies only by ticker, so a topic like "ai" is not representable upstream and is a keyword rule over the headline and summary here.

Case-insensitive search over headlines and summaries of the current board.

Bitcoin only, beta. One timeframe's analysis, its OHLCV candles and the support, resistance and FBB band levels the analysis lists. `granularity` names the bars: 2H, 6H, 8H and 12H are built from TSP's 1h bars, and 1m, 3m, 5m and 15m have levels only, no bars (the finest bars the Signals run takes from TSP are 30m); `note` says so. `bars` defaults to the newest 50 and accepts up to 200. The newest bar can carry a partial flag: when true it was read before it closed, ingested_at says when (null if not known), and its high, low and close are as of then. When TSP also marks it live, it has no ingested_at: its high, low and close are a live sample TSP took at as_of (null if not known; extremes is TSP's word for how its high and low were taken, sampled or null), and volume_ingested_at says when its volume was recorded (null if not known). A bar TSP makes from its sample before that interval's bar is recorded is left out, so a series ends at its newest recorded bar: partial is true there when TSP moved its flag from it onto the bar it made, or left it unlabelled with its volume recorded before it closed. Levels, not a trade plan. Not investment advice.

Bitcoin only, beta. The BTC derivatives market: funding, open interest, liquidations, long/short, taker volume, ETF flows, options max pain and a fear and greed index (0 to 100, 100 = extreme greed), each with its own observed time and cadence. Open interest comes for one contract (`instrument`, its venue in `exchange`) and for several together: `contracts` lists the perpetual contracts in that one figure, `contracts_unlisted` counts any leg it does not name, and `includes_single_contract` says whether the one contract is among them; a venue not shown here is null. `flow.retail` is the all-accounts long/short ratio (the global account ratio, by number of accounts), set against the top-trader position ratio. A pending ETF day has a null flow, never 0, whatever TSP's own word for it (tsp_status: reported, provisional or null). Clock verdicts (cadence_check, open_bucket.closed, stopped_refreshing, max pain's nearest expiry) are judged at the moment of the answer. By default the latest reading of every metric is complete and the back-history beside it is the newest 3 points per series; the answer's `history` block reports how many older points were left out. Set `history` to true for the full stored series. Not investment advice.

Bitcoin only, beta. When the BTC signals snapshot was made, how old its newest value is, whether it is degraded and why, and every source with its own observed time and whether it was carried from an earlier run. Also the beta and not-advice notice. It is the only one of the signals tools that is purely about the age of the data.

Bitcoin only, beta. The BTC spot price with the change TSP reports beside it (`spot.delta_pct`, in percent; TSP does not say over what period), the trend, direction and confidence on each of thirteen timeframes with their consensus, the bull, base and bear cases with their probabilities for three horizons (TSP's own labels for them, `most_probable_arm` and `dominant_case`, are defined in mcw-signals-thesis), and support and resistance by how often they recur. Direction and confidence only: no entry, target or stop is published. Every value carries its own observed time, and `freshness` says how old the snapshot is. Each horizon's inputs are judged at the moment of the answer, each series against its own cadence (judged_on[].cadence_check, and stale per input family; ETF flows by two days since their row was written, as stopped_refreshing judges them). Not investment advice.

Bitcoin only, beta. The bull, base and bear cases for three horizons (scalp, short and long), as the Signals page's Thesis tab shows them. For each horizon: `most_probable_arm`, TSP's own most probable case (bull, base, bear or none; null when not known), with `probabilities_tie` true when two or more cases share the highest probability, and TSP's pick among them then its own tie-break (null when a probability is missing); `dominant_case`, TSP's own label: the highest-probability case with price levels (bull, base, bear or none; null when not known; the levels themselves are not published), so it can name a different case from the most probable one; the aggregate confidence; and each case's action, its probability, whether it has a signal for its direction (`has_signal`) and whether TSP marks it degenerate. Each horizon's inputs are judged at the moment of the answer, each series against its own cadence (judged_on[].cadence_check, and stale per input family; ETF flows by two days since their row was written). `horizon_timeframes` names the timeframe analyses each horizon is built from, and `timeframes` holds their rows, each with its own analyzed_at. Direction and probability only: no entry, target or stop is published, nor any of TSP's reasoning text. Not investment advice.