ATLAS · API

Probability analytics
for US equities.
One endpoint.

Add Monte Carlo probability, earnings-blended forecasts, and AI analyst narratives to your platform — without building a quant team.

4,500+
US equities
<100ms
cached p95
513K
probabilities/night
POWERED BY
Professional-tier market dataFederal Reserve (FRED · 15+ indicators)Business-tier news feedAI synthesis layer
 1{ 2  "summary": { 3    "macroRegime": "BEARISH", 4    "macroLabel": "mu x0.8 sigma x1.1", 5    "pathsSimulated": 10000, 6    "diffusionModel": "Student-t (df=5)" 7  }, 8  "assets": [{ 9    "symbol": "AAPL",10    "price": 273.17,11    "verdict": "Roughly flat",12    "expectedReturnPct": 0.14,13    "probabilityHit": 0.4790,14    "baseProbabilityHit": 0.0151,15    "riskProfile": {16      "pLose5": 0.2634,17      "pLose10": 0.0999,18      "pLose20": 0.003519    },20    "earnings": {21      "inHorizon": true,22      "blendFormula": "(1-0.79) x 1.5% + 0.79 x 60.0%",23      "probabilityAfterEarnings": 0.479024    },25    "aiEnrichment": {26      "analystVerdict": "Apple is in a steady-state product cycle...",27      "catalysts": ["iPhone 18 Pro upgrade cycle", ...],28      "risks": ["CEO transition in September...", ...]29    }30  }]31}
At a glance

The endpoint, summarized.

Everything a developer needs to know in one block. The full reference doc is shared after request access.

Endpoint
POST /api/simulate-full
Universe
~4,500 US-listed equities (NYSE + Nasdaq)
Sectors
All 11 GICS, with per-sector news enrichment
Horizons
7, 15, 21, 30, 60, 90 trading days
Target returns
19 discrete targets per horizon, from 1% to 500%
Loss thresholds
5%, 10%, 20%, 30%, 50% (precomputed per stock)
Refresh cadence
Nightly end-of-day batch
Monte Carlo
10,000 paths per stock per horizon
Diffusion model
Student-t with 5 degrees of freedom (fat-tailed)
Response format
Single structured JSON, all derivations precomputed
Two modes
Cached probability (sub-100ms) and AI-enriched
Why partners pick ATLAS

Stop building probability.
Start shipping it.

We run 10,000-path Monte Carlo on every covered stock, every night, across 6 horizons and 19 targets. You hit one endpoint and get the full distribution back as JSON. The math is solved. You ship features.

What you get

Three things, end to end.

Full coverage

4,500+ US equities, refreshed nightly

  • NYSE + Nasdaq across all 11 GICS sectors
  • 6 horizons (7, 15, 21, 30, 60, 90 trading days)
  • 19 target returns from 1% to 500%
  • 5 loss thresholds (5/10/20/30/50%)

Two response modes

Pick speed or pick depth — same endpoint

  • Cached: <100ms p95, pure probability + technical
  • Enriched: AI analyst verdict, catalysts, risks
  • Toggle via skipNews flag in request body
  • Cached fallback if AI layer fails

Real transparency

Every input, every multiplier, every blend exposed

  • Raw σ and µ shown alongside macro-adjusted versions
  • Earnings blend formula returned as a string
  • News multiplier and sentiment scores in response
  • Hard-risk warning flag for regulatory/fraud events
Real example

Request.
Response.

AAPL · +20% target · 30 trading days · enriched mode. Captured from the production endpoint.

 1curl -X POST https://api.atlas-stocks.com/api/simulate-full \ 2  -H "Content-Type: application/json" \ 3  -d '{ 4    "symbols": ["AAPL"], 5    "targetReturnPct": 20, 6    "horizonDays": 30, 7    "skipNews": false 8  }'
What's inside the math

Four layers.
All exposed.

Tap any layer to see the formula

Downside, exposed

The riskProfile object.
Five thresholds, monotonic.

Every response includes a complete downside distribution at five drawdown thresholds. Always monotonically non-increasing — if pLose20 is 0.5%, pLose30 cannot exceed it.

Real AAPL · 30-day horizon · +20% target
pLose5
26.34%
Chance of losing 5% or more
pLose10
9.99%
Chance of losing 10% or more
pLose20
0.35%
Chance of losing 20% or more
pLose30
~0%
Chance of losing 30% or more
pLose50
0%
Chance of losing 50% or more
Render this as a step chart or bar chart. End users find loss visualizations more informative than single-number summaries — they reveal the shape of the tail.
Pre-labeled UI blocks

keyDrivers.
Drop-in cards.

Every response includes a keyDrivers object with four UI-ready blocks. Each has a label, a direction tag (positive / neutral / negative), and a descriptive blurb. Drop them into a four-up card layout — no client-side classification logic needed.

riskLevel

Risk level

Classifies the stock's volatility against typical equity benchmarks. Derived from annualizedSigmaPct.

σ > 35%
High
20% < σ ≤ 35%
Moderate
σ ≤ 20%
Low
trend

Trend

Directional bias of the expected return. Derived from expectedReturnPct.

expected > 1%
Upward
−1% ≤ exp ≤ 1%
Flat
expected < −1%
Downward
earningsImpact

Earnings impact

Binary — earnings either fall inside the horizon or they don't. When they do, dir is set by the historical hit rate.

No earnings in horizon
Not scheduled
In horizon · hist > 15%
Upcoming
In horizon · hist ≤ 15%
Upcoming
newsImpact

News impact

Reflects the news sentiment adjustment applied to the probability. Derived from globalScore.

globalScore > 0.1
Positive
−0.1 ≤ score ≤ 0.1
Neutral
globalScore < −0.1
Slightly negative
Recommended integration

Two calls.
Instant UI.

The pattern that minimizes latency and maximizes perceived responsiveness. Most clients render the cached response in under 100ms, then progressively populate the analyst narrative when it arrives.

01

Fire cached call immediately

skipNews: true

When a user opens a stock page, request the cached response. It returns in under 100ms. Render probability, riskProfile, verdict, keyDrivers, full transparency block. The UI is fully populated.

02

Fire enriched call in parallel

skipNews: false

Simultaneously, request the AI-enriched response. On a cold cache it can take 8–12s. While it loads, show a subtle skeleton in the "Analyst view" section. Don't block the rest of the UI.

03

Populate when enriched arrives

data.aiEnrichment

When the enriched response lands, fill in analystVerdict, catalysts, risks, newsContext. Subsequent stock opens within 4–5 hours hit cache and return both calls in <100ms.

Operational reality

Latency. Errors. Limits.

Everything you need to know before integrating.

Latency profile

Mode
p50
p95
p99
Cached
40ms
90ms
180ms
Enriched · cold cache
9s
14s
22s
Enriched · within 4–5h cache
50ms
110ms
210ms
Error / unsupported symbol
15ms
30ms

Error model

400
Bad Request
Malformed JSON, missing fields, unsupported horizon
401
Unauthorized
Missing or invalid API key
429
Too Many Requests
Includes X-RateLimit-* headers
500
Internal Server Error
Logged with trace ID for our side
503
Service Unavailable
Cached responses still serve
Partial success: if one symbol in a batch fails, that one returns an error object. The rest succeed normally.
Who uses this

Built for partners.

If you serve retail investors and you don't already have a quant team, ATLAS is the fastest way to ship probability features.

Robo-advisors

Risk-aware allocation suggestions

  • Pull riskProfile per holding for client portfolios
  • Surface earnings-blended probability on tilts
  • Use macroRegime to gate aggressive recommendations

Brokers + trading apps

Pre-trade probability context

  • Show hit probability before order entry
  • Render full distribution on stock detail pages
  • Flag hard-risk warnings in real time

Investment platforms

Education + research surfaces

  • Embed analyst verdict in research feeds
  • Drive newsletter content from catalyst/risk lists
  • Power "what could happen" widgets at zero compute cost
Pricing

Custom by volume and mode.

No public price list. Cached-mode and enriched-mode have different unit economics; we size pricing to your traffic and your tier mix. Fill the form and we'll come back with a number within 48 hours.

Get access

Tell us what you're
building.

Fill this out and we'll come back within 48 hours with a sandbox key and a pricing proposal.

We receive everything at info@atlas-stocks.com and reply within 1 business day.

Probability,
as a service.

You ship features. We solve the math.

Request access
ATLAS – See the Odds Before You Invest