Building with the API? Open the developer docs
Finmagine API & MCP — free tier to test • Pro/Max for full access • REST + MCP
Master the Momentum Setup Workflow through a complete learning overview and interactive flashcards
Not every recipe in agent-recipes.json chains one call's output into the next. The Momentum Setup Workflow is the simplest recipe in the whole catalog — 4 independent read calls with genuinely empty input_map on every step — and understanding why that's the correct design, not a missing feature, is exactly what this guide covers. It also covers a real, honest engineering decision: why the US version deliberately ships with only 3 of the 4 steps.
Click any card to reveal the answer. Use the search box to find a specific topic.
Four independent, synchronous read calls — no depends_on, no LLM, no BYO key, no async polling. Run all 4, then cross-reference the results yourself (or let an agent cross-reference them). Total runtime: under a minute.
| Property | Value |
|---|---|
| Tier | API Pro |
| Async? | No — all synchronous reads |
| LLM required? | No |
| Quota | None — these are ordinary read calls, not gated resources like Committee/Preset runs |
| Estimated runtime | Under 1 minute (4 synchronous read calls) |
get_market_breadth — last 30 days of advances/declines, % of stocks above 10/20/40-day moving averages, and Nifty close. This is a position-sizing gate, checked before looking at any individual stock:
| 40-day MA breadth | Regime | Implication |
|---|---|---|
| > 65% | Broad advance | Full-size entries appropriate |
| 50–65% | Selective market | Reduce position size |
| < 40% | Correction | Hold cash |
| Sharp single-day drop to <30% advancers | Distribution | A warning signal independent of the trailing breadth level |
get_rs_changes — stocks whose RS rating has improved the most over N days. Stocks early in an RS improvement cycle often lead the next phase of the move, before price itself has caught up. An RS delta greater than 20 points in 20 days is considered strong acceleration.
get_market_momentum — the current RS leaderboard (RS ≥ 90 = elite). This is where a stock already is, as opposed to Step 2's "where it's heading."
get_conviction_ranking — stocks passing the most simultaneous Finmagine scans (Stage 2, VCP, near-52-week-high, RS momentum, PEAD, and more). A stock appearing in 4 or more scans simultaneously is worth investigating regardless of sector.
This recipe's output_contract states it plainly: "Four independent lists to cross-reference by hand/agent-judgment — stocks appearing across get_rs_changes, get_market_momentum, AND get_conviction_ranking simultaneously are the highest-confluence candidates."
This is a genuinely different pattern from the Screen → Committee recipe: there, step outputs mechanically feed the next step's request via input_map. Here, all 4 outputs land side by side, and the actual analytical work — finding the overlap — happens after all 4 calls return, not during them.
The recipe's failure_fallback field reads: "None of these calls have external dependencies — a failure here is a genuine API/auth problem, not a data-availability issue."
us_momentum_setup_partial ships with 3 of the 4 steps — Breadth, Momentum, and Conviction — using the US equivalents (us_get_market_breadth, us_market_momentum, us_get_conviction_ranking), all at Free tier. The missing piece is an equivalent of get_rs_changes.
note field states it directly: "US has no equivalent of India's get_rs_changes ('RS improving before price') — blocked on a missing dated RS-rating history pipeline for US equities, not just a missing endpoint. Do not substitute a same-day cross-sectional RS comparison as a stand-in; it is a materially different signal from 'RS improved over time,' and a prior review explicitly rejected shipping a fake substitute under this name. Deferred to a future pass once the pipeline exists."
A same-day cross-sectional RS comparison (ranking today's stocks by RS against each other, right now) and a genuine "RS improved over N days" trend signal answer different questions. The first tells you who's strong today; the second tells you who's getting strong — the specific "early in an RS improvement cycle" signal India's version provides. Building the dated history needed to compute the second kind of signal for US equities is real infrastructure work that hadn't been done yet — and shipping the first kind under the same name would have quietly misrepresented what the US recipe actually measures.
status field is explicitly partial rather than complete, the gap is documented in plain language in the note field, and the substitute that was considered and rejected is named specifically — so nobody building on this recipe discovers the gap the hard way, mid-integration.
Both the India and US versions of this recipe carry the same explicit guardrail field: "Signal-generation workflow, not investment advice."
Every number this workflow returns — breadth percentages, RS deltas, scan-overlap counts — is a real, computed market statistic. What they mean for any specific decision is left entirely to you or your agent's own judgment; the workflow surfaces signal, not a recommendation.
The final guide in this 12-part series covers the other named workflow built on the endpoints from the previous guide:
us_market_trader_scan's scan types together for a tighter, technically-driven shortlistSee the API & Agent Integration hub for the full 12-guide roadmap.
Finmagine gives you 30+ computed financial ratios, sector benchmarks, FII/DII flows, the Finmagine Score, and AI-powered analysis — all in one place.