Fetching Forward Curves

Fetch forward rate curves for pricing, forecasting, and building interest rate models.

The /forward_curve endpoint returns a series of forward rates over time. Use this for dashboards, pricing models, cash flow projections, and interest rate forecasts.

---

Basic Example

import requests

url = "https://api.bluegamma.io/v1/forward_curve"
headers = {"x-api-key": "your_api_key"}

params = {
    "index": "SOFR",
    "start_date": "0D",
    "end_date": "10Y",
    "tenor": "3M",
    "frequency": "3M"
}

response = requests.get(url, headers=headers, params=params)
data = response.json()
curve = data["curve"]

Response:

{
  "start_date": "2025-12-17",
  "end_date": "2035-12-17",
  "index": "SOFR",
  "valuation_time": "2025-12-17T10:22:04.246666",
  "tenor": "3M",
  "frequency": "3M",
  "curve": [
    {"start_date": "2025-12-17", "end_date": "2026-03-17", "forward_rate": 4.31},
    {"start_date": "2026-03-17", "end_date": "2026-06-17", "forward_rate": 3.89},
    {"start_date": "2026-06-17", "end_date": "2026-09-17", "forward_rate": 3.65},
    {"start_date": "2026-09-17", "end_date": "2026-12-17", "forward_rate": 3.52}
  ],
  "timestamp": "2025-12-17T10:21:29"
}

Field	Description
curve	Array of forward rate points
forward_rate	The forward rate as a percentage
start_date / end_date	The period the forward rate covers
valuation_time	When the curve was calculated
timestamp	When the market data was captured

---

Parameters

Parameter	Required	Description
index	Yes	Interest rate index (e.g., SOFR, SONIA, 6M EURIBOR)
start_date	Yes	Start of the curve, relative (0D, 1M, 1Y) or absolute (2025-12-17)
end_date	Yes	End of the curve
tenor	Yes	Length of each forward period (e.g., 3M, 6M)
frequency	Yes	Spacing between curve points (e.g., 3M, 1Y)
valuation_time	No	Historical timestamp (ISO 8601) for reproducibility

---

When to Use Forward Curves

Use Case	Example
Cash flow projection	Project floating rate payments on a loan or swap
Dashboard / UI	Display the market's rate expectations
Scenario analysis	Compare current curve to historical snapshots
Model calibration	Feed forward rates into a pricing model

SOFR forward curve: now vs 1 year ago

The chart shows how rate expectations have shifted over the past year:

• Dec 2024 (grey): Relatively flat curve around 3.75–3.95%

• Dec 2025 (blue): Sharp dip to ~3.1% at 1Y (rate cuts priced in), then steepening to 4.4% at 10Y

Use valuation_time to compare curves at different points in time.

---

Curve Access Patterns: Fetch vs Store

When integrating forward curves, you can either fetch on-demand or store snapshots locally. The right choice depends on your workflow.

Fetch on Demand

Best for flexibility and fresh data:

• Live dashboards and UIs

• Ad-hoc pricing with user-defined date ranges

• One-off analysis

# Just call the API when you need it
response = requests.get(url, headers=headers, params=params)
curve = response.json()["curve"]

Store Snapshots

Best for speed, reproducibility, and audit trails:

• Monthly/quarterly valuations

• Batch processing that queries the same curves repeatedly

• Any workflow where you need to prove what curve you used

# Fetch and store with full context
response = requests.get(url, headers=headers, params=params)
data = response.json()

db.insert("curve_snapshots", {
    "valuation_time": data["valuation_time"],
    "timestamp": data["timestamp"],
    "index": data["index"],
    "curve": data["curve"]
})

Quick Comparison

	Fetch on Demand	Store Snapshots
Setup	Simple	Requires storage
Speed	API latency	Instant
Reproducibility	Requires logging	Built-in
Best for	Real-time, ad-hoc	Periodic, auditable

---

Reproducibility with valuation_time

For reproducible valuations, use valuation_time to lock in the market state at a specific point in time:

params = {
    "index": "SOFR",
    "start_date": "0D",
    "end_date": "10Y",
    "tenor": "3M",
    "frequency": "3M",
    "valuation_time": "2024-06-30T16:00:00Z"  # ISO 8601 format
}

response = requests.get(url, headers=headers, params=params)

Without valuation_time, you get the latest data. With it, you get the curve exactly as it was at that timestamp—essential for reproducing valuations months later.

---

Hybrid Approach

Many workflows combine both patterns—store snapshots for fixed valuation dates, fetch on demand for everything else:

def get_curve(index: str, valuation_time: str = None):
    # Check local cache first
    if valuation_time:
        stored = db.query(
            "SELECT curve FROM snapshots WHERE index = ? AND valuation_time = ?",
            [index, valuation_time]
        )
        if stored:
            return stored["curve"]
    
    # Fetch from API
    params = {
        "index": index,
        "start_date": "0D",
        "end_date": "10Y",
        "tenor": "3M",
        "frequency": "3M"
    }
    if valuation_time:
        params["valuation_time"] = valuation_time
    
    response = requests.get(url, headers=headers, params=params)
    return response.json()["curve"]

---

Storage Schema

If you're storing curves, here's a suggested schema:

Column	Type	Description
valuation_time	TIMESTAMP	When the curve was calculated
timestamp	TIMESTAMP	Market data timestamp
index	VARCHAR	e.g., "SOFR", "6M EURIBOR"
curve	JSON	Array of {start_date, end_date, forward_rate}
created_at	TIMESTAMP	When you captured it

---

Supported Indices

Index	Currency	Description
SOFR	USD	Secured Overnight Financing Rate
SONIA	GBP	Sterling Overnight Index Average
ESTR	EUR	Euro Short-Term Rate
6M EURIBOR	EUR	6-Month Euro Interbank Offered Rate
CORRA	CAD	Canadian Overnight Repo Rate Average

For the full list, see Available Indices.

---

Related Endpoints

If you need...	Use
A single forward rate	/forward_rate
Zero rates for discounting	/zero_rate
Discount factors	/discount_factor
The underlying swap curve	/swap_curve

---

Need an API key? 📩 [email protected] | 📅 Book a call