sygnldata — market intelligence for trading agents

Market regime, execution-cost, bar-QC and backtest-audit tools for agents.

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

    What data it sees

    Do you need an account

    No: the server works without sign-in

    Market regime, execution-cost, bar-QC and backtest-audit tools for agents. Pay per call via x402.

    How to connect

    How to connect sygnldata — market intelligence for trading agents to Claude

    The server address to paste into Claude settings:

    https://api.sygnliq.com/mcp

    1. Open Claude (claude.ai or the desktop app).
    2. Go to Settings → Connectors.
    3. Click Add custom connector.
    4. Paste the server address copied below into Remote MCP server URL and click Add.
    5. In a chat, click + → Connectors and switch the new connector on.

    Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

    Authorization

    After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

    How to connect sygnldata — market intelligence for trading agents to ChatGPT

    The server address to paste into ChatGPT settings:

    https://api.sygnliq.com/mcp

    1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
    2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
    3. Open Settings → Connectors and click Create.
    4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
    5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
    6. In a new chat open + → Apps/Connectors and enable the connector.

    OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

    Authorization

    On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

    How to connect sygnldata — market intelligence for trading agents to Cursor

    The server address to paste into Cursor settings:

    https://api.sygnliq.com/mcp

    Fastest: click Open in Cursor below and confirm the prompt.

    Manually:

    1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
    2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
    3. Save the file. The server appears in the MCP list; authorize it there if asked.

    Authorization

    If the service needs sign-in, an authorize button appears next to the server in the MCP list.

    Server tool list (21)

    Raw names from tools/list. Only developers need these.

    get_regime_sampleFREE: 24h-delayed market regime snapshot (taste of the live paid feed).
    get_pricingFREE: price list, payment network, and payTo address for all paid tools.
    get_regime_currentLive market regime: STABLE/VOLATILE + confidence + position_scalar ($0.005). Validated variance detector: VOLATILE predicts ~2x forward volatility with ~90% persistence. Use for position sizing / risk gating, NOT direction.
    get_regime_historyHistorical regime snapshots, 5-min cadence since 2026-02, 417k+ rows ($0.02). start/end are ISO timestamps; limit <= 10000.
    estimate_execution_costRealistic execution cost in bps for an order ($0.01). Calibrated to measured live fills (SPY/QQQ 1.5bps round-trip floor), with square-root size impact. order_type: limit_nbbo | market.
    check_bar_qualityForensic QC on YOUR OHLCV bars before you backtest on them ($0.05). Detects impossible bars, bad prints, unadjusted splits, gaps, stale data. bars: [{timestamp, open, high, low, close, volume}, ...] (5..100000 rows)
    audit_backtestForensic audit of a backtest trade ledger ($5.00). Impossible-fill detection against the real tape, cost-adjusted expectancy, overfitting checks, verdict + grade. trades_csv: CSV text with columns like symbol, side, entry_time, exit_time, entry_price, exit_price, qty.
    trade_preflightPre-trade decision gate for AI trading agents ($0.01). Deterministic ALLOW / REDUCE / REQUIRE_CONFIRMATION / BLOCK with reason codes, an order-hash-bound approved size, 60s expiry, and an EIP-191-signed attestation. Composes the live regime detector, the measured execution-cost oracle, and YOUR policy (max_notional_usd, max_cost_bps, daily_loss_limit_usd, confirm_above_usd, symbol_allowlist, ...). order: {symbol, side, notional_usd, order_type}. evidence: {audit_score} from audit_backtest. Not investment advice.
    sygnl47_manifestFREE: SYGNL/47 service manifest — method, latency tiers, prices, inventory. Read this before calling any paid sygnl47_* tool. Describes how events are detected and verified (live-vs-replay acoustic fingerprinting, multi-source corroboration, linguistic guards, abnormal-move attribution), the full price ladder, and the CURRENT INVENTORY stated plainly. Inventory today is thin and we would rather you knew: 2 events, both reconstructed from archived broadcast audio of a single speech and labelled source: replay, and ZERO confirmed live outcomes. This manifest contains no event data — it is a description of the service, not a free sample. There is no free tier of the data itself, only cheaper (older) tiers.
    sygnl47_latest_eventMost recent verified POTUS statement event, no delay ($2.00). RETURNS: the single newest CONFIRMED observation — timestamp, entity and ticker named, a verbatim snippet capped at 180 characters, the verification block (live vs replay, how many independent feeds corroborated it, consensus confidence) and the market block (mark, benchmark, abnormal return at 1/5/15/30 minutes, instrument minus benchmark). GUARANTEED MAX STALENESS: 0 seconds of deliberate delay. Real freshness is bounded by publisher sync and reported in every response as sync.feed_age_seconds — measured, never assumed. PRICE: $2.00 USDC. Deliberately premium: this is verified, hard-to- replicate data and we would rather learn what one real buyer pays than what a hundred free riders sample. NOT A PREDICTION. This reports what was said and what happened afterwards. No edge claim, no win rate, no recommendation. If no event exists the call returns 404 and is NOT settled — you are not charged for silence.
    sygnl47_events_liveVerified POTUS statement events with no delay applied ($2.00). RETURNS: newest-first list of CONFIRMED observations, same schema as every other tier: verification block (live vs replay via acoustic fingerprinting, independent source count) and market block (abnormal return vs benchmark). GUARANTEED MAX STALENESS: 0 seconds. PRICE: $2.00 USDC — the top of the ladder, because recency is the only thing the price buys. NOT A PREDICTION: verified observation of a public statement and the price action that followed. Zero confirmed live outcomes to date. since: ISO-8601 lower bound on ts_utc (narrows only). limit: 1..500.
    sygnl47_events_1mVerified POTUS statement events, nothing newer than 60 seconds ($0.75). RETURNS: identical schema, identical fields and identical verification block to the live tier. Nothing is stripped — only recency is. GUARANTEED MAX STALENESS: 60 seconds. The cutoff is enforced server-side on ts_utc and bound to this priced route, not to any argument you send. PRICE: $0.75 USDC. NOT A PREDICTION: verified observation, no edge claim.
    sygnl47_events_15mVerified POTUS statement events, nothing newer than 15 minutes ($0.25). RETURNS: the same complete payload as the live tier, including the full verification and market-attribution blocks. GUARANTEED MAX STALENESS: 900 seconds. PRICE: $0.25 USDC. Appropriate for attribution and research rather than reaction. NOT A PREDICTION: verified observation, no edge claim.
    sygnl47_events_archiveVerified POTUS statement events at T+24h and older ($0.10). RETURNS: the full historical record, same complete schema as every other tier — the archive is not a sample and not a degraded copy. GUARANTEED MAX STALENESS: 86400 seconds. PRICE: $0.10 USDC, the cheapest rung — deliberately kept low enough that a price-sensitive agent still has something it can transact on. INVENTORY, HONESTLY: 2 events today, both reconstructed from archived broadcast audio of one speech (2026-07-27, GM Milford Proving Ground) and labelled source: replay with sources: 1 and replay_checked: false. Two fires from one speech are not two independent observations. NOT A PREDICTION: verified observation, no edge claim, zero confirmed live outcomes.
    sygnl47_event_detailFull detail for one SYGNL/47 event by id ($2.00). RETURNS: the complete event — verbatim snippet (180-char cap), verification block (live vs replay, liveness reason, independent source count, consensus confidence) and market attribution (mark, benchmark, abnormal return by horizon). GUARANTEED MAX STALENESS: none — this serves events of any age including the newest, which is why it is priced at the live tier rather than the archive tier. PRICE: $2.00 USDC. An unknown event_id returns 404 and is NOT settled. NOT A PREDICTION: verified observation of a public statement. event_id example: sygnl47-20260727-192342-GM
    sygnl47_entitiesTracked entity gazetteer plus today's guard state ($0.25). RETURNS: every entity SYGNL/47 listens for with its disambiguation rules attached — a country only counts when a policy instrument is nearby, "intel" does not count next to "agency", "apple" does not count next to "pie" — and, per ticker, whether it is BLOCKED right now by the earnings and liquidity guards, with the reason string. WHY IT CHANGES: a name being watched and a name being actionable today are different facts. An earnings print inside the holding window destroys move attribution, so those names are refused outright. Liquidity below $10M 20-session average dollar volume is a hard reject. A guard that could not check reports UNKNOWN and never silently passes. GUARANTEED MAX STALENESS: guard state is resolved at publish time; the response carries guards_evaluated_utc and sync.feed_age_seconds so you can judge it yourself. PRICE: $0.25 USDC. NOT A PREDICTION: this is a reference and constraint dataset, not a recommendation to trade any name it lists as allowed.
    sygnl47_feedsBroadcast feed buffer-depth map and source health ($0.25). RETURNS: per live news feed, the measured connect-buffer depth — how many seconds of already-buffered audio the source dumps the instant you connect — plus transport kind and uptime observation counts. The measurement behind it found continuous Icecast radio buffers 2-5x shallower than segmented video HLS (NPR 2.7s vs Reuters 38.3s), which is a genuinely useful fact for anyone building a low-latency listening path. STATED PLAINLY: this is buffer depth, NOT standing latency. Standing latency is a different number, it requires a simulcast we have not captured, and the cross-feed correlation available today scores below its own acceptance threshold — so no per-feed lag figure is published. We would rather return a null than a confident wrong number. PRICE: $0.25 USDC. NOT A PREDICTION: infrastructure measurement only.
    sygnl47_subscription_plansFREE: prepaid-pass plans for SYGNL/47 (the subscription catalog). x402 is per-call by design. A subscription here is a PREPAID TIME-BOXED PASS on the same rails: one payment buys unmetered access to one tier for one window and returns a signed bearer token, which you then send as X-SYGNL47-TOKEN instead of paying per call. Lists each plan with its tier, duration, price and the exact routes it authorizes. Read this before calling any sygnl47_subscribe_* tool.
    sygnl47_subscribe_live_24hBuy a 24-hour unmetered pass to the SYGNL/47 live tier ($50.00). RETURNS: {token, sub_id, tier, expires_utc, max_freshness_seconds, authorized_routes}. Send the token as header X-SYGNL47-TOKEN on any authorized route and no per-call payment is charged until it expires. STORE THE TOKEN — verification is stateless by design, so we cannot reissue it. WHAT IT DOES NOT BUY: fresher data. The pass grants unmetered access to the live tier, and the server-side ts_utc cutoff applies to pass holders exactly as it does to per-call buyers. An expired token falls back to a normal 402 challenge, never to a free serve. PRICE: $50.00 USDC, versus $2.00 per live call. NOT A PREDICTION: this buys access to a verified observation feed with zero confirmed live outcomes, whose current inventory is 2 replay-sourced events.
    sygnl47_subscribe_live_30dBuy a 30-day unmetered pass to the SYGNL/47 live tier ($750.00). Same contract as the 24-hour pass over a 30-day window: signed bearer token, unmetered access to the live tier, no widening of freshness beyond that tier, stateless verification, and a normal 402 fallback once it expires. PRICE: $750.00 USDC. NOT A PREDICTION: verified observation feed, no edge or win-rate claim, zero confirmed live outcomes to date. At this price you should read sygnl47_manifest first — it states the inventory plainly, and the inventory is currently 2 events reconstructed from archived audio.
    sygnl47_subscribe_archive_30dBuy a 30-day unmetered pass to the SYGNL/47 archive tier ($100.00). RETURNS: a signed bearer token granting unmetered access to the archive tier (T+24h and older) for 30 days. The archive payload is complete, not a sample — same schema and same verification block as the live tier, only older. IMPORTANT: an archive pass does NOT authorize the live, 1m or 15m routes. Using it there returns a normal 402 with an explanation, because a pass may never widen freshness beyond the tier it was bought for. PRICE: $100.00 USDC. NOT A PREDICTION: verified observation feed, no edge claim.