---
title: "Predict Das"
method: POST
path: "/api/v1/predict/das"
---

# Predict Das

`POST /api/v1/predict/das`

## Request

- Base URL: `https://www.deepmapai.com`
- URL: `https://www.deepmapai.com/api/v1/predict/das`
- Auth: none declared

DAS analysis of YOUR record: FK, events, interferometry, spectral + ML outliers.

Runs das.analyze.full_analysis on the record the caller supplies in `data`
(samples x channels, >= 10 s, sample interval `dt`, channel spacing `dx`), and
flags channels whose spectral features are outliers (Isolation Forest at the
caller's `contamination`). An outlier channel is not by itself a void, a leak or a
coupling fault; it is a channel to inspect.

No `data` -> HTTP 422. Until 2026-10-14 this endpoint ignored the caller and
analysed server-generated random numbers (np.random.randn), unlabelled.

Pricing (unchanged): $2K-$10K/mo per monitored fiber segment.

## Headers

- `x-api-key` string

## Request body

- DASPredictRequest: DAS analysis request -- computed ONLY from the caller's `data`. Until 2026-10-14 there was no `data` field: the endpoints analysed server-generated noise (and, for earthquake-detect / classify, an injected synthetic event).
  - `data` array[], nullable: REQUIRED. DAS record, shape (samples, channels). The server generates no input.
    - number[]
  - `n_samples` integer, nullable: Ignored -- the shape is taken from `data`.
  - `n_channels` integer, nullable: Ignored -- the shape is taken from `data`.
  - `dt` number: Sample interval of `data` (s)
  - `dx` number: Channel spacing (m)
  - `contamination` number: Expected anomaly fraction (Isolation Forest)

## Response `200`

Successful Response

- unknown

## Other responses

- `422`: Validation Error

---

[API](https://skmtc.dev/deepmapai/apis/deepmap-ai.md) · [All operations](https://skmtc.dev/deepmapai/apis/deepmap-ai/llms.txt) · [OpenAPI document](https://skmtc.dev/deepmapai/apis/deepmap-ai/revisions/dab995d977d0?raw)