SuperTrend K — 股市起漲K線 API integration & OpenData services

1. Authorized account & session module

Login flow analysis (CMoney SSO + device fingerprint + session token refresh), token vault, multi-account rotation. Used to keep a long-running data agent connected without re-prompting the end user, and to bind one CMoney account to multiple downstream consumers.

2. Rising-signal stream

Captures every "AI 自動盯盤" alert — uptrend breakout, range break, V-shape reversal — plus the underlying indicator scores. Delivered as a websocket or webhook with stock_id, signal_type, score, fired_at, and last_close. Used for fast-fill discretionary desks and for backtesting signal decay.

3. Broker chip-flow connector

Daily ingest of branch-level buy/sell concentration, top-15 broker buyers and sellers, and your imported broker holdings. Used for "main force tracking" dashboards, position-risk reports, and triggering rebalancing logic when chip concentration crosses a threshold.

4. Long-vs-short multi-factor pull

Per-ticker JSON snapshot combining chip flow, price-volume, technical, and fundamental factors — the same four pillars the app surfaces in its long-short table. Used to feed a single "bias score" into proprietary screeners or PM dashboards.

5. Strong "mad-dog flow" screener API

One endpoint that runs the 9-indicator strength × 6-indicator volume screener and returns the ranked candidate list with the matching reason codes. Used in morning-briefing emails and to seed an automated paper-trading account on 股市大富翁 for forward testing.

6. Discussion-area & sentiment hooks

Authorized scraping of public threads in CMoney's investment discussion area, scoped per stock_id. Returns thread title, author hash, post timestamp, and a precomputed sentiment label. Used as an auxiliary feature in event-driven models and for catching unusual post-volume spikes ahead of price moves.

Scenario A — Quant desk feature pipeline

Context: A Taipei-based quant fund already buys TWSE tick from Fugle Market Data but wants a derived breakout-signal feature without rebuilding the indicator stack.

Data & APIs: Rising-signal stream + technical indicator bundle + branch-level chip concentration, normalized into a unified signals_intraday Parquet partition.

OpenData mapping: Sits next to TWSE OpenAPI and FinMind feeds in the team's lakehouse, accessed through one OpenFinance-style gateway with per-key rate limits and audit logs.

Scenario B — Family-office portfolio mirror

Context: A family office holds positions across two brokers and wants the SuperTrend K "broker inventory" monitor to mirror the same book the PMs see in their internal tool.

Data & APIs: Authorized broker-holdings import + rising-signal webhook scoped to held tickers + Excel export endpoint for monthly review.

OpenData mapping: Acts as a private OpenBanking-style "consented data feed" — the user authorizes once, the studio's gateway proxies and de-duplicates.

Scenario C — Roboadvisor model overlay

Context: A retail roboadvisor wants to add a "momentum overlay" to its passive ETF baskets without licensing a full third-party signal platform.

Data & APIs: Mad-dog flow screener + long/short multi-factor bias score, polled nightly. Output flows into the model's tilt-weights endpoint.

OpenData mapping: The studio supplies a thin OpenFinance-style adapter (OAuth2 client credentials, JSON schemas) so the roboadvisor's compliance team can review one contract instead of two.

Scenario D — Line / Telegram alert bot

Context: A Taiwan-based paid newsletter wants premium subscribers to receive Line alerts the moment the SuperTrend K engine flags a "mad-dog flow" candidate from their curated list.

Data & APIs: Rising-signal webhook filtered by a watchlist of ~120 symbols + bot-friendly REST endpoint that returns the human-readable reason code.

OpenData mapping: Treated as a one-way OpenData export — no PII flows back, only signal metadata and the symbol-level commentary.

Scenario E — Cross-platform research dashboard

Context: A research analyst keeps watchlists in 富果 Fugle, 富邦 e點通, and SuperTrend K, and needs a single source of truth for "where am I currently exposed and which of my names just fired a signal."

Data & APIs: Watchlist round-trip + rising-signal stream + a thin reconciliation service that diffs symbols across the three providers nightly.

OpenData mapping: Implemented as an OpenFinance-style aggregator with per-source consent records and a scoped audit log, ready for the analyst's compliance officer.

A. Authorized session bootstrap (pseudocode)

POST /api/v1/supertrendk/session
Content-Type: application/json
X-Studio-Client: ccwork-prod

{
  "account_ref": "user_8c41",
  "device_fp": "tw-android-9f02...",
  "scope": ["signals.read", "broker.read", "watchlist.rw"]
}

200 OK
{
  "session_id": "sess_01HV...",
  "access_token": "eyJhbGciOi...",
  "expires_in": 3300,
  "refresh_after": 2400,
  "consent_record": "pdpa-2026-04-29-tw"
}

B. Rising-signal webhook payload

POST {YOUR_ENDPOINT}/supertrendk/signal
X-Signature: sha256=...
Content-Type: application/json

{
  "fired_at": "2026-04-29T05:01:18+08:00",
  "stock_id": "2330",
  "market": "TWSE",
  "signal_type": "uptrend_breakout",
  "score": 0.83,
  "indicators": {
    "strength_hits": 7,
    "volume_hits": 5
  },
  "last_close": 945.0,
  "reason_codes": ["MA20_break", "vol3x", "chip_in"]
}

C. Mad-dog flow screener pull

GET /api/v1/supertrendk/screener/maddogflow
  ?as_of=2026-04-29
  &market=TWSE,OTC
  &min_score=0.6
  &page_size=50
Authorization: Bearer <ACCESS_TOKEN>

200 OK
{
  "as_of": "2026-04-29",
  "items": [
    {"stock_id":"3017","score":0.91,"reasons":["KD_golden","vol_burst"]},
    {"stock_id":"6488","score":0.87,"reasons":["range_break","chip_concentration"]}
  ],
  "next_cursor": null,
  "audit_id": "aud_01HV..."
}

# Errors are surfaced as RFC 7807 problem+json with
# stable codes: rate_limited, scope_denied, upstream_throttled.

D. Excel / JSON statement export

POST /api/v1/supertrendk/export
{
  "account_id": "user_8c41",
  "from_date": "2026-01-01",
  "to_date":   "2026-04-29",
  "datasets":  ["signals", "broker_holdings"],
  "format":    "xlsx"
}

200 OK
{
  "file_url": "https://files.example.com/...sig",
  "expires_at": "2026-04-29T08:00:00Z",
  "row_counts": {"signals": 1284, "broker_holdings": 312}
}

Deliverables checklist

• OpenAPI 3.1 / Swagger spec for every endpoint above
• Protocol & auth-flow report (CMoney login chain, token lifecycle, rate-limit map)
• Runnable source: ingest agent + gateway + sample consumers (Python or Node.js)
• Schema definitions for signals, broker holdings, screener output, watchlists
• Postman collection and a one-shot smoke-test script
• Compliance pack — PDPA mapping, FSC outsourcing notes, retention defaults
• Operational runbook: token rotation, error taxonomy, on-call playbook

Engagement models

Source-code delivery (from $300): we ship the runnable repo and docs, you host. Pay after delivery upon satisfaction. Fits teams with internal SRE who prefer to own the runtime.

Pay-per-call API (no upfront): use our hosted endpoints, pay only for the calls you make. Fits roboadvisors, newsletters, and prototype-stage projects that want zero capex.

Hybrid is allowed — start hosted, migrate to source-code delivery once volume is stable.

About us

We are an independent studio focused on mobile-app interface analysis and authorized financial API integration. The team has shipped connectors for Taiwan, Hong Kong, Japan, India, and Brazil markets, mixing reverse-engineering rigor with practical compliance work. We know the difference between a quant team that needs Parquet and a Line-bot operator that needs a webhook, and we scope accordingly.

Contact

Send the target app name, the data points you actually need, and any sandbox / account credentials you want us to use. We reply with a scoping memo and a fixed-fee quote within two business days.

Open contact page

Engagement workflow

1. Scope confirmation — which screens, which data, which target system. (1–2 business days)
2. Protocol analysis & API design — OpenAPI spec, auth chain, rate-limit map. (2–5 business days)
3. Build & internal validation — ingest agent, gateway, consumers, smoke tests. (3–8 business days)
4. Docs, samples, runbook — Postman, retention defaults, on-call notes. (1–2 business days)
5. Handover & first production drop — typically 5–15 business days end-to-end. Third-party approvals (e.g. FSC outsourcing review) may extend the timeline.

FAQ

Do you need my CMoney credentials?

Can you cover both TWSE and OTC tickers?

What if CMoney ships an app update that breaks the integration?

Can the same gateway proxy other Taiwan stock apps?